idnits 2.17.1 draft-ietf-rmt-pi-norm-10.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- ** It looks like you're using RFC 3978 boilerplate. You should update this to the boilerplate described in the IETF Trust License Policy document (see https://trustee.ietf.org/license-info), which is required now. -- Found old boilerplate from RFC 3978, Section 5.5 on line 3526. ** Found boilerplate matching RFC 3978, Section 5.4, paragraph 1 (on line 3518), which is fine, but *also* found old RFC 2026, Section 10.4C, paragraph 1 text on line 35. ** The document claims conformance with section 10 of RFC 2026, but uses some RFC 3978/3979 boilerplate. As RFC 3978/3979 replaces section 10 of RFC 2026, you should not claim conformance with it if you have changed to using RFC 3978/3979 boilerplate. ** The document seems to lack an RFC 3978 Section 5.1 IPR Disclosure Acknowledgement. ** This document has an original RFC 3978 Section 5.4 Copyright Line, instead of the newer IETF Trust Copyright according to RFC 4748. ** The document seems to lack an RFC 3978 Section 5.4 Reference to BCP 78 -- however, there's a paragraph with a matching beginning. Boilerplate error? ** This document has an original RFC 3978 Section 5.5 Disclaimer, instead of the newer disclaimer which includes the IETF Trust according to RFC 4748. ** The document seems to lack an RFC 3979 Section 5, para. 1 IPR Disclosure Acknowledgement. ** The document seems to lack an RFC 3979 Section 5, para. 2 IPR Disclosure Acknowledgement. ** The document seems to lack an RFC 3979 Section 5, para. 3 IPR Disclosure Invitation. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- == No 'Intended status' indicated for this document; assuming Proposed Standard == The page length should not exceed 58 lines per page, but there was 70 longer pages, the longest (page 1) being 63 lines Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- ** There are 113 instances of too long lines in the document, the longest one being 40 characters in excess of 72. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the RFC 3978 Section 5.4 Copyright Line does not match the current year == Line 306 has weird spacing: '...inguish betwe...' == Line 1053 has weird spacing: '... source or...' == Line 1713 has weird spacing: '...rs have recei...' == Line 1978 has weird spacing: '...al from when ...' == Line 2442 has weird spacing: '...length given ...' == (4 more instances...) == Using lowercase 'not' together with uppercase 'MUST', 'SHALL', 'SHOULD', or 'RECOMMENDED' is not an accepted usage according to RFC 2119. Please use uppercase 'NOT' together with RFC 2119 keywords (if that is what you mean). Found 'SHALL not' in this paragraph: The NORM_CMD(SQUELCH) command is transmitted in response to outdated or invalid NORM_NACK content received by the sender. Invalid NORM_NACK content consists of repair requests for NormObjects for which the sender is unable or unwilling to provide repair. This includes repair requests for outdated objects, aborted objects, or those objects which the sender previously transmitted marked with the NORM_FLAG_UNRELIABLE flag. This command indicates to receivers what content is available for repair, thus serving as a description of the sender's current "repair window". Receivers SHALL not generate repair requests for content identified as invalid by a NORM_CMD(SQUELCH). -- The document seems to lack a disclaimer for pre-RFC5378 work, but may have content which was first submitted before 10 November 2008. If you have contacted all the original authors and they are all willing to grant the BCP78 rights to the IETF Trust, then this is fine, and you can ignore this comment. If not, you may need to add the pre-RFC5378 disclaimer. (See the Legal Provisions document at https://trustee.ietf.org/license-info for more information.) -- The document date (July 2004) is 7224 days in the past. Is this intentional? Checking references for intended status: Proposed Standard ---------------------------------------------------------------------------- (See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) ** Downref: Normative reference to an Informational RFC: RFC 3269 (ref. '1') ** Downref: Normative reference to an Experimental draft: draft-ietf-rmt-bb-norm (ref. '4') ** Obsolete normative reference: RFC 3452 (ref. '5') (Obsoleted by RFC 5052, RFC 5445) ** Obsolete normative reference: RFC 2434 (ref. '6') (Obsoleted by RFC 5226) -- Obsolete informational reference (is this intentional?): RFC 2327 (ref. '7') (Obsoleted by RFC 4566) -- Obsolete informational reference (is this intentional?): RFC 1889 (ref. '18') (Obsoleted by RFC 3550) -- Obsolete informational reference (is this intentional?): RFC 2401 (ref. '22') (Obsoleted by RFC 4301) Summary: 15 errors (**), 0 flaws (~~), 10 warnings (==), 6 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 RMT Working Group B. Adamson/NRL 3 INTERNET-DRAFT C. Bormann/Tellique 4 draft-ietf-rmt-pi-norm-10 M. Handley/ACIRI 5 Expires: January 2005 J. Macker/NRL 6 July 2004 8 NACK-Oriented Reliable Multicast Protocol (NORM) 10 Status of this Memo 12 This document is an Internet-Draft and is in full conformance with all 13 provisions of Section 10 of RFC2026. By submitting this Internet-Draft, 14 we certify that any applicable patent or other IPR claims of which we 15 are aware have been disclosed, and any of which we become aware will be 16 disclosed, in accordance with RFC 3668. 18 Internet-Drafts are working documents of the Internet Engineering Task 19 Force (IETF), its areas, and its working groups. Note that other groups 20 may also distribute working documents as Internet-Drafts. 22 Internet-Drafts are draft documents valid for a maximum of six months 23 and may be updated, replaced, or obsoleted by other documents at any 24 time. It is inappropriate to use Internet-Drafts as reference material 25 or to cite them other than as "work in progress." 27 The list of current Internet-Drafts can be accessed at 28 http://www.ietf.org/ietf/1id-abstracts.txt 30 The list of Internet-Draft Shadow Directories can be accessed at 31 http://www.ietf.org/shadow.html. 33 Copyright Notice 35 Copyright (C) The Internet Society (2004). All Rights Reserved. 37 Abstract 39 This document describes the messages and procedures of the Negative- 40 acknowledgment (NACK) Oriented Reliable Multicast (NORM) protocol. This 41 protocol is designed to provide end-to-end reliable transport of bulk 42 data objects or streams over generic IP multicast routing and forwarding 43 services. NORM uses a selective, negative acknowledgment mechanism for 44 transport reliability and offers additional protocol mechanisms to allow 45 for operation with minimal "a priori" coordination among senders and 46 receivers. A congestion control scheme is specified to allow the NORM 47 protocol fairly share available network bandwidth with other transport 48 protocols such as Transmission Control Protocol (TCP). It is capable of 49 operating with both reciprocal multicast routing among senders and 50 receivers and with asymmetric connectivity (possibly a unicast return 51 path) between the senders and receivers. The protocol offers a number 52 of features to allow different types of applications or possibly other 53 higher level transport protocols to utilize its service in different 54 The protocol leverages the use of FEC-based repair and other IETF 55 reliable multicast transport (RMT) building blocks in its design. 57 Table of Contents 59 1. Introduction and Applicability. . . . . . . . . . . . . . . . . . 3 60 1.1. NORM Delivery Service Model. . . . . . . . . . . . . . . . . . 3 61 1.2. NORM Scalability . . . . . . . . . . . . . . . . . . . . . . . 5 62 1.3. Environmental Requirements and Considerations. . . . . . . . . 6 63 2. Architecture Definition . . . . . . . . . . . . . . . . . . . . . 6 64 2.1. Protocol Operation Overview. . . . . . . . . . . . . . . . . . 8 65 2.2. Protocol Building Blocks . . . . . . . . . . . . . . . . . . . 9 66 2.3. Design Tradeoffs . . . . . . . . . . . . . . . . . . . . . . . 9 67 3. Conformance Statement . . . . . . . . . . . . . . . . . . . . . . 10 68 4. Message Formats . . . . . . . . . . . . . . . . . . . . . . . . . 12 69 4.1. NORM Common Message Header and Extensions. . . . . . . . . . . 13 70 4.2. Sender Messages. . . . . . . . . . . . . . . . . . . . . . . . 15 71 4.2.1. NORM_DATA Message . . . . . . . . . . . . . . . . . . . . . 15 72 4.2.2. NORM_INFO Message . . . . . . . . . . . . . . . . . . . . . 22 73 4.2.3. NORM_CMD Messages . . . . . . . . . . . . . . . . . . . . . 24 74 4.3. Receiver Messages. . . . . . . . . . . . . . . . . . . . . . . 39 75 4.3.1. NORM_NACK Message . . . . . . . . . . . . . . . . . . . . . 39 76 4.3.2. NORM_ACK Message. . . . . . . . . . . . . . . . . . . . . . 45 77 4.4. General Purpose Messages . . . . . . . . . . . . . . . . . . . 46 78 4.4.1. NORM_REPORT Message . . . . . . . . . . . . . . . . . . . . 46 79 5. Detailed Protocol Operation . . . . . . . . . . . . . . . . . . . 47 80 5.1. Sender Initialization and Transmission . . . . . . . . . . . . 48 81 5.1.1. Object Segmentation Algorithm . . . . . . . . . . . . . . . 49 82 5.2. Receiver Initialization and Reception. . . . . . . . . . . . . 51 83 5.3. Receiver NACK Procedure. . . . . . . . . . . . . . . . . . . . 51 84 5.4. Sender NACK Processing and Response. . . . . . . . . . . . . . 53 85 5.4.1. Sender Repair State Aggregation . . . . . . . . . . . . . . 53 86 5.4.2. Sender FEC Repair Transmission Strategy . . . . . . . . . . 54 87 5.4.3. Sender NORM_CMD(SQUELCH) Generation . . . . . . . . . . . . 55 88 5.4.4. Sender NORM_CMD(REPAIR_ADV) Generation. . . . . . . . . . . 55 89 5.5. Additional Protocol Mechanisms . . . . . . . . . . . . . . . . 56 90 5.5.1. Greatest Round-trip Time Collection . . . . . . . . . . . . 56 91 5.5.2. NORM Congestion Control Operation . . . . . . . . . . . . . 57 92 5.5.3. NORM Positive Acknowledgment Procedure. . . . . . . . . . . 64 93 5.5.4. Group Size Estimate . . . . . . . . . . . . . . . . . . . . 66 94 6. Security Considerations . . . . . . . . . . . . . . . . . . . . . 66 95 7. IANA Considerations . . . . . . . . . . . . . . . . . . . . . . . 67 96 8. Suggested Use . . . . . . . . . . . . . . . . . . . . . . . . . . 67 97 9. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . 68 98 10. References . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 99 10.1. Normative References. . . . . . . . . . . . . . . . . . . . . 68 100 10.2. Informative References. . . . . . . . . . . . . . . . . . . . 68 101 11. Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 69 102 12. Full Copyright Statement . . . . . . . . . . . . . . . . . . . . 70 103 1. Introduction and Applicability 105 The Negative-acknowledgment (NACK) Oriented Reliable Multicast (NORM) 106 protocol is designed to provide reliable transport of data from one or 107 more sender(s) to a group of receivers over an IP multicast network. 108 The primary design goals of NORM are to provide efficient, scalable, and 109 robust bulk data (e.g., computer files, transmission of persistent data) 110 transfer across possibly heterogeneous IP networks and topologies. The 111 NORM protocol design provides support for distributed multicast session 112 participation with minimal coordination among senders and receivers. 113 NORM allows senders and receivers to dynamically join and leave 114 multicast sessions at will with minimal overhead for control information 115 and timing synchronization among participants. To accommodate this 116 capability, NORM protocol message headers contain some common 117 information allowing receivers to easily synchronize to senders 118 throughout the lifetime of a reliable multicast session. NORM is 119 designed to be self-adapting to a wide range of dynamic network 120 conditions with little or no pre-configuration. The protocol is 121 purposely designed to be tolerant of inaccurate timing estimations or 122 lossy conditions that may occur in many networks including mobile and 123 wireless. The protocol is also designed to exhibit convergence and 124 efficient operation even in situations of heavy packet loss and large 125 queuing or transmission delays. 127 This document is a product of the IETF RMT WG and follows the guidelines 128 provided in RFC 3269 [1]. The key words "MUST", "MUST NOT", 129 "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", 130 "MAY", and "OPTIONAL" in this document are to be interpreted as 131 described in BCP 14, RFC 2119 [2]. 133 1.1. NORM Delivery Service Model 135 A NORM protocol instance (NormSession) is defined within the context of 136 participants communicating connectionless (e.g., Internet Protocol (IP) 137 or User Datagram Protocol (UDP)) packets over a network using pre- 138 determined addresses and host port numbers. Generally, the participants 139 exchange packets using an IP multicast group address, but unicast 140 transport may also be established or applied as an adjunct to multicast 141 delivery. In the case of multicast, the participating NormNodes will 142 communicate using a common IP multicast group address and port number 143 that has been chosen via means outside the context of the given 144 NormSession. Other IETF data format and protocol standards exist that 145 may be applied to describe and convey the required "a priori" 146 information for a specific NormSession (e.g., Session Description 147 Protocol (SDP) [7], Session Announcement Protocol (SAP) [8], etc). 149 The NORM protocol design is principally driven by the assumption of a 150 single sender transmitting bulk data content to a group of receivers. 151 However, the protocol MAY operate with multiple senders within the 152 context of a single NormSession. In initial implementations of this 153 protocol, it is anticipated that multiple senders will transmit 154 independent of one another and receivers will maintain state as 155 necessary for each sender. However, in future versions of NORM, it is 156 that some aspects of protocol operation (e.g., round-trip time 157 collection) may provide for alternate modes allowing more efficient 158 performance for applications requiring multiple senders. 160 NORM provides for three types of bulk data content objects (NormObjects) 161 to be reliably transported. These types include: 163 1 static computer memory data content (NORM_OBJECT_DATA type), 165 2) computer storage files (NORM_OBJECT_FILE type), and 167 3) non-finite streams of continuous data content 168 (NORM_OBJECT_STREAM type). 170 The distinction between NORM_OBJECT_DATA and NORM_OBJECT_FILE is simply 171 to provide a "hint" to receivers in NormSessions serving multiple types 172 of content as to what type of storage should be allocated for received 173 content (i.e. memory or file storage). Other than that distinction, the 174 two are identical, providing for reliable transport of finite (but 175 potentially very large) units of content. These static data and file 176 services are anticipated to be useful for multicast-based cache 177 applications with the ability to reliably provide transmission of large 178 quantities of static data. Other types of static data/file delivery 179 services might make use of these transport object types, too. The use 180 of the NORM_OBJECT_STREAM type is at the application's discretion and 181 could be used to carry static data or file content also. The NORM 182 reliable stream service opens up additional possibilities such as 183 serialized reliable messaging or other unbounded, perhaps dynamically 184 produced content. The NORM_OBJECT_STREAM provides for reliable 185 transport analogous to that of the Transmission Control Protocol (TCP), 186 although NORM receivers will be able to begin receiving stream content 187 at any point in time. The applicability of this feature will depend 188 upon the application. 190 The NORM protocol also allows for a small amount of "out-of-band" data 191 (sent as NORM_INFO messages) to be attached to the data content objects 192 transmitted by the sender. This readily-available "out-of-band" data 193 allows multicast receivers to quickly and efficiently determine the 194 nature of the corresponding data, file, or stream bulk content being 195 transmitted. This allows application-level control of the receiver 196 node's participation in the current transport activity. This also 197 allows the protocol to be flexible with minimal pre-coordination among 198 senders and receivers. The NORM_INFO content is designed to be atomic 199 in that its size MUST fit into the payload portion of a single NORM 200 message. 202 NORM does _not_ provide for global or application-level identification 203 of data content within in its message headers. Note the NORM_INFO out- 204 of-band data mechanism could be leveraged by the application for this 205 purpose if desired, or identification could alternatively be embedded 206 within the data content. NORM does identify transmitted content 207 with transport identifiers that are applicable only while the sender is 208 transmitting and/or repairing the given object. These transport data 209 content identifiers (NormTransportIds) are assigned in a monotonically 210 increasing fashion by each NORM sender during the course of a 211 NormSession. Each sender maintains its NormTransportId assignments 212 independently so that individual NormObjects may be uniquely identified 213 during transport with the concatenation of the sender session-unique 214 identifier (NormNodeId) and the assigned NormTransportId. The 215 NormTransportIds are assigned from a large, but fixed, numeric space in 216 increasing order and may be reassigned during long-lived sessions. The 217 NORM protocol provides mechanisms so that the sender application may 218 terminate transmission of data content and inform the group of this in 219 an efficient manner. Other similar protocol control mechanisms (e.g., 220 session termination, receiver synchronization, etc) are specified so 221 that reliable multicast application variants may construct different, 222 complete bulk transfer communication models to meet their goals. 224 To summarize, the NORM protocol provides reliable transport of different 225 types of data content (including potentially mixed types). The senders 226 enqueue and transmit bulk content in the form of static data or files 227 and/or non-finite, ongoing stream types. NORM senders provide for 228 repair transmission of data and/or FEC content in response to NACK 229 messages received from the receiver group. Mechanisms for "out-of-band" 230 information and other transport control mechanisms are specified for use 231 by applications to form complete reliable multicast solutions for 232 different purposes. 234 1.2. NORM Scalability 236 Group communication scalability requirements lead to adaptation of 237 negative acknowledgment (NACK) based protocol schemes when feedback for 238 reliability is required [9]. NORM is a protocol centered around the 239 use of selective NACKs to request repairs of missing data. NORM 240 provides for the use of packet-level forward error correction (FEC) 241 techniques for efficient multicast repair and optional proactive 242 transmission robustness [10]. FEC-based repair can be used to 243 greatly reduce the quantity of reliable multicast repair requests and 244 repair transmissions [11] in a NACK-oriented protocol. The principal 245 factor in NORM scalability is the volume of feedback traffic generated 246 by the receiver set to facilitate reliability and congestion control. 247 NORM uses probabilistic suppression of redundant feedback based on 248 exponentially distributed random backoff timers. The performance of 249 this type of suppression relative to other techniques is described in 250 [12]. NORM dynamically measures the group's roundtrip timing status to 251 set its suppression and other protocol timers. This allows NORM to 252 scale well while maintaining reliable data delivery transport with low 253 latency relative to the network topology over which it is operating. 255 Feedback messages can be either multicast to the group at large or sent 256 via unicast routing to the sender. In the case of unicast feedback, the 257 sender "advertises" the feedback state to the group to facilitate 258 feedback suppression. In typical Internet environments, it is expected 259 that the NORM protocol will readily scale to group sizes on the order of 260 of thousands of receivers. A study of the quantity of feedback for this 261 type of protocol is described in [13]. NORM is able to operate with a 262 smaller amount of feedback than a single TCP connection, even with 263 relatively large numbers of receivers. Thus, depending upon the network 264 topology, it is possible that NORM may scale to larger group sizes. 265 With respect to computer resource usage, the NORM protocol does _not_ 266 require that state be kept on all receivers in the group. NORM senders 267 maintain state only for receivers providing explicit congestion control 268 feedback. NORM receivers must maintain state for each active sender. 269 This may constrain the number of simultaneous senders in some uses of 270 NORM. 272 1.3. Environmental Requirements and Considerations 274 All of the environmental requirements and considerations that apply to 275 the RMT NORM Building Block [4] and the RMT FEC Building Block [5] 276 also apply to the NORM protocol. 278 The NORM protocol SHALL be capable of operating in an end-to-end fashion 279 with no assistance from intermediate systems beyond basic IP multicast 280 group management, routing, and forwarding services. While the 281 techniques utilized in NORM are principally applicable to "flat" end-to- 282 end IP multicast topologies, they could also be applied in the sub- 283 levels of hierarchical (e.g., tree-based) multicast distribution if so 284 desired. NORM can make use of reciprocal (among senders and receivers) 285 multicast communication under the Any-Source Multicast (ASM) model 286 defined in RFC 1112 [3], but SHALL also be capable of scalable 287 operation in asymmetric topologies such as Source Specific Multicast 288 (SSM) [14] where there may only be unicast routing service from the 289 receivers to the sender(s). 291 NORM is compatible with IPv4 and IPv6. Additionally, NORM may be used 292 with networks employing Network Address Translation (NAT) providing the 293 NAT device supports IP multicast and/or can cache UDP traffic source 294 port numbers for remapping feedback traffic from receivers to the 295 sender(s). 297 2. Architecture Definition 299 A NormSession is comprised of participants (NormNodes) acting as senders 300 and/or receivers. NORM senders transmit data content in the form of 301 NormObjects to the session destination address and the NORM receivers 302 attempt to reliably receive the transmitted content using negative 303 acknowledgments to request repair. Each NormNode within a NormSession 304 is assumed to have a preselected unique 32-bit identifier (NormNodeId). 305 NormNodes MUST have uniquely assigned identifiers within a single 306 NormSession to distinguish between possible multiple senders and to 307 distinguish feedback information from different receivers. There are 308 two reserved NormNodeId values. A value of 0x00000000 is considered an 309 invalid NormNodeId value and a value of 0xffffffff is a "wildcard" 310 NormNodeId. While the protocol does not preclude multiple sender nodes 311 concurrently transmitting within the context of a single NORM session 312 (i.e. many- to-many operation), any type of interactive coordination 313 NORM senders is assumed to be controlled by the application or higher 314 protocol layer. There are some optional mechanisms specified in this 315 document that can be leveraged for such application layer coordination. 317 As previously noted, NORM allows for reliable transmission of three 318 different basic types of data content. The first type is 319 NORM_OBJECT_DATA, which is used for static, persistent blocks of data 320 content maintained in the sender's application memory storage. The 321 second type is NORM_OBJECT_FILE, which corresponds to data stored in the 322 sender's non-volatile file system. The NORM_OBJECT_DATA and 323 NORM_OBJECT_FILE types both represent "NormObjects" of finite but 324 potentially very large size. The third type of data content is 325 NORM_OBJECT_STREAM, which corresponds to an ongoing transmission of 326 undefined length. This is analogous to the reliable stream service 327 provide by TCP for unicast data transport. The format of the stream 328 content is application-defined and may be byte or message oriented. The 329 NORM protocol provides for "flushing" of the stream to expedite delivery 330 or possibly enforce application message boundaries. NORM protocol 331 implementations may offer either (or both) in-order delivery of the 332 stream data to the receive application or out-of-order (more immediate) 333 delivery of received segments of the stream to the receiver application. 334 In either case, NORM sender and receiver implementations provide 335 buffering to facilitate repair of the stream as it is transported. 337 All NormObjects are logically segmented into FEC coding blocks and 338 symbols for transmission by the sender. In NORM, an FEC encoding symbol 339 directly corresponds to the payload of NORM_DATA messages or "segment". 340 Note that when systematic FEC codes are used, the payload of NORM_DATA 341 messages sent for the first portion of a FEC encoding block are source 342 symbols (actual segments of original user data), while the remaining 343 symbols for the block consist of parity symbols generated by FEC 344 encoding. These parity symbols are generally sent in response to repair 345 requests, but some number may be sent proactively at the end each 346 encoding block to increase the robustness of transmission. When non- 347 systematic FEC codes are used, all symbols sent consist of FEC encoding 348 parity content. In this case, the receiver must receive a sufficient 349 number of symbols to reconstruct (via FEC decoding) the original user 350 data for the given block. In this document, the terms "symbol" and 351 "segment" are used interchangeably. 353 Transmitted NormObjects are temporarily yet uniquely identified within 354 the NormSession context using the given sender's NormNodeId, 355 NormInstanceId, and a temporary NormObjectTransportId. Depending upon 356 the implementation, individual NORM senders may manage their 357 NormInstanceIds independently, or a common NormInstanceId may be agreed 358 upon for all participating nodes within a session if needed as a session 359 identifier. NORM NormObjectTransportId data content identifiers are 360 sender-assigned and applicable and valid only during a NormObject's 361 actual _transport_ (i.e. for as long as the sender is transmitting and 362 providing repair of the indicated NormObject). For a long-lived 363 session, the NormObjectTransportId field can wrap and previously-used 364 identifiers may be re-used. Note that globally unique identification of 365 transported data content is not provided by NORM and, if required, must 366 managed by the NORM application. The individual segments or symbols of 367 the NormObject are further identified with FEC payload identifiers which 368 include coding block and symbol identifiers. These are discussed in 369 detail later in this document. 371 2.1. Protocol Operation Overview 373 A NORM sender primarily generates messages of type NORM_DATA. These 374 messages carry original data segments or FEC symbols and repair 375 segments/symbols for the bulk data/file or stream NormObjects being 376 transferred. By default, redundant FEC symbols are sent only in 377 response to receiver repair requests (NACKs) and thus normally little or 378 no additional transmission overhead is imposed due to FEC encoding. 379 However, the NORM implementation MAY be optionally configured to 380 proactively transmit some amount of redundant FEC symbols along with the 381 original content to potentially enhance performance (e.g., improved 382 delay) at the cost of additional transmission overhead. This option may 383 be sensible for certain network conditions and can allow for robust, 384 asymmetric multicast (e.g., unidirectional routing, satellite, cable) 385 [15] with reduced receiver feedback, or, in some cases, no feedback. 387 A sender message of type NORM_INFO is also defined and is used to carry 388 OPTIONAL "out-of-band" context information for a given transport object. 389 A single NORM_INFO message can be associated with a NormObject. Because 390 of its atomic nature, missing NORM_INFO messages can be NACKed and 391 repaired with a slightly lower delay process than NORM's general FEC- 392 encoded data content. NORM_INFO may serve special purposes for some bulk 393 transfer, reliable multicast applications where receivers join the group 394 mid-stream and need to ascertain contextual information on the current 395 content being transmitted. The NACK process for NORM_INFO will be 396 described later. When the NORM_INFO message type is used, its 397 transmission should precede transmission of any NORM_DATA message for 398 the associated NormObject. 400 The sender also generates messages of type NORM_CMD to assist in certain 401 protocol operations such as congestion control, end-of-transmission 402 flushing, round trip time estimation, receiver synchronization, and 403 optional positive acknowledgment requests or application defined 404 commands. The transmission of NORM_CMD messages from the sender is 405 accomplished by one of three different procedures. These procedures 406 are: single, best effort unreliable transmission of the command; 407 repeated redundant transmissions of the command; and positively- 408 acknowledged commands. The transmission technique used for a given 409 command depends upon the function of the command. Several core commands 410 are defined for basic protocol operation. Additionally, implementations 411 MAY wish to consider providing the OPTIONAL application-defined commands 412 that can take advantage of the transmission methodologies available for 413 commands. This allows for application-level session management 414 mechanisms that can make use of information available to the underlying 415 NORM protocol engine (e.g., round-trip timing, transmission rate, etc). 417 NORM receivers generate messages of type NORM_NACK or NORM_ACK in 418 response to transmissions of data and commands from a sender. The 419 messages are generated to request repair of detected data transmission 420 losses. Receivers generally detect losses by tracking the sequence of 421 transmission from a sender. Sequencing information is embedded in the 422 transmitted data packets and end-of-transmission commands from the 423 sender. NORM_ACK messages are generated in response to certain commands 424 transmitted by the sender. In the general (and most scalable) protocol 425 mode, NORM_ACK messages are sent only in response to congestion control 426 commands from the sender. The feedback volume of these congestion 427 control NORM_ACK messages is controlled using the same timer-based 428 probabilistic suppression techniques as for NORM_NACK messages to avoid 429 feedback implosion. In order to meet potential application requirements 430 for positive acknowledgment from receivers, other NORM_ACK messages are 431 defined and available for use. All sender and receiver transmissions 432 are subject to rate control governed by a peak transmission rate set for 433 each participant by the application. This can be used to limit the 434 quantity of multicast data transmitted by the group. When NORM's 435 congestion control algorithm is enabled the rate for senders is 436 automatically adjusted. In some networks, it may be desirable to 437 establish minimum and maximum bounds for the rate adjustment depending 438 upon the application even when dynamic congestion control is enabled. 439 However, in the case of the general Internet, congestion control policy 440 SHALL be observed which is compatible with coexistent TCP flows. 442 2.2. Protocol Building Blocks 444 The operation of the NORM protocol is based primarily upon the concepts 445 presented in the Nack-Oriented Reliable Multicast (NORM) Building Block 446 document [4]. This includes the basic NORM architecture and the data 447 transmission, repair, and feedback strategies discussed in that 448 document. Additional reliable multicast building blocks are applied in 449 creating the full NORM protocol instantiation [16]. NORM also makes 450 use of Forward Error Correction encoding techniques for repair messaging 451 and optional transmission robustness as described in [10]. NORM uses 452 the FEC Payload ID as specified by the FEC Building Block Document [5]. 453 Additionally, for congestion control, this document includes a baseline 454 congestion control mechanism (NORM-CC) based on the TCP-Friendly 455 Multicast Congestion Control (TFMCC) scheme described in [19]. 457 2.3. Design Tradeoffs 459 While the various features of NORM are designed to provide some measure 460 of general purpose utility, it is important to emphasize the 461 understanding that "no one size fits all" in the reliable multicast 462 transport arena. There are numerous engineering tradeoffs involved in 463 reliable multicast transport design and this requires an increased 464 awareness of application and network architecture considerations. 465 Performance requirements affecting design can include: group size, 466 heterogeneity (e.g., capacity and/or delay), asymmetric delivery, data 467 ordering, delivery delay, group dynamics, mobility, congestion control, 468 and transport across low capacity connections. NORM contains various 469 parameters to accommodate many of these differing requirements. The 470 NORM protocol and its mechanisms MAY be applied in multicast 471 applications outside of bulk data transfer, but there is an assumed 472 of bulk transfer transport service that drives the trade-offs that 473 determine the scalability and performance described in this document. 475 The ability of NORM to provide reliable data delivery is also governed 476 by any buffer constraints of the sender and receiver applications. NORM 477 protocol implementations SHOULD be designed to operate with the greatest 478 efficiency and robustness possible within application-defined buffer 479 constraints. Buffer requirements for reliability, as always, are a 480 function of the delay-bandwidth product of the network topology. NORM 481 performs best when allowed more buffering resources than typical point- 482 to-point transport protocols. This is because NORM feedback suppression 483 is based upon randomly-delayed transmissions from the receiver set, 484 rather than immediately transmitted feedback. There are definitive 485 tradeoffs between buffer utilization, group size scalability, and 486 efficiency of performance. Large buffer sizes allow the NORM protocol 487 to perform most efficiently in large delay-bandwidth topologies and 488 allow for longer feedback suppression backoff timeouts. This yields 489 improved group size scalability. NORM can operate with reduced 490 buffering but at a cost of decreased efficiency (lower relative goodput) 491 and reduced group size scalability. 493 3. Conformance Statement 495 This Protocol Instantiation document, in conjunction with the RMT 496 Building Block documents of [4] and [5], completely specifies a 497 working reliable multicast transport protocol that conforms to the 498 requirements described in RFC 2357 [17]. 500 This document specifies the following message types and mechanisms which 501 are REQUIRED in complying NORM protocol implementations: 503 +---------------------+-----------------------------------------------+ 504 | Message Type | Purpose | 505 +---------------------+-----------------------------------------------+ 506 |NORM_DATA | Sender message for application data | 507 | | transmission. Implementations must support | 508 | | at least one of the NORM_OBJECT_DATA, | 509 | | NORM_OBJECT_FILE, or NORM_OBJECT_STREAM | 510 | | delivery services. The use of the NORM FEC | 511 | | Object Transmission Information header | 512 | | extension is OPTIONAL with NORM_DATA | 513 | | messages. | 514 +---------------------+-----------------------------------------------+ 515 |NORM_CMD(FLUSH) | Sender command to excite receivers for repair | 516 | | requests in lieu of ongoing NORM_DATA | 517 | | transmissions. Note the use of the | 518 | | NORM_CMD(FLUSH) for positive acknowledgment | 519 | | of data receipt is OPTIONAL. | 520 +---------------------+-----------------------------------------------+ 521 |NORM_CMD(SQUELCH) | Sender command to advertise its current valid | 522 | | repair window in response to invalid requests | 523 | | for repair. | 524 +---------------------+-----------------------------------------------+ 525 |NORM_CMD(REPAIR_ADV) | Sender command to advertise current repair | 526 | | (and congestion control state) to group when | 527 | | unicast feedback messages are detected. Used | 528 | | to control/suppress excessive receiver | 529 | | feedback in asymmetric multicast topologies. | 530 +---------------------+-----------------------------------------------+ 531 |NORM_CMD(CC) | Sender command used in collection of round | 532 | | trip timing and congestion control status | 533 | | from group (This may be OPTIONAL if | 534 | | alternative congestion control mechanism and | 535 | | round trip timing collection is used). | 536 +---------------------+-----------------------------------------------+ 537 |NORM_NACK | Receiver message used to request repair of | 538 | | missing transmitted content. | 539 +---------------------+-----------------------------------------------+ 540 |NORM_ACK | Receiver message used to proactively provide | 541 | | feedback for congestion control purposes. | 542 | | Also used with the OPTIONAL NORM Positive | 543 | | Acknowledgment Process. | 544 +---------------------+-----------------------------------------------+ 546 This document also describes the following message types and associated 547 mechanisms which are OPTIONAL for complying NORM protocol 548 implementations: 550 +-----------------------+-----------------------------------------------+ 551 | Message Type | Purpose | 552 +-----------------------+-----------------------------------------------+ 553 |NORM_INFO | Sender message for providing ancillary | 554 | | context information associated with NORM | 555 | | transport objects. The use of the NORM FEC | 556 | | Object Transmission Information header | 557 | | extension is OPTIONAL with NORM_INFO | 558 | | messages. | 559 +-----------------------+-----------------------------------------------+ 560 |NORM_CMD(EOT) | Sender command to indicate it has reach end- | 561 | | of-transmission and will no longer respond to | 562 | | repair requests. | 563 +-----------------------+-----------------------------------------------+ 564 |NORM_CMD(ACK_REQ) | Sender command to support application- | 565 | | defined, positively acknowledged commands | 566 | | sent outside of the context of the bulk data | 567 | | content being transmitted. The NORM Positive | 568 | | Acknowledgment Procedure associated with this | 569 | | message type is OPTIONAL. | 570 +-----------------------+-----------------------------------------------+ 571 |NORM_CMD(APPLICATION) | Sender command containing application-defined | 572 | | commands sent outside of the context of the | 573 | | bulk data content being transmitted. | 574 +-----------------------+-----------------------------------------------+ 575 |NORM_REPORT | Optional message type reserved for | 576 | | experimental implementations of the NORM | 577 | | protocol. | 578 +-----------------------+-----------------------------------------------+ 580 4. Message Formats 582 As mentioned in Section 2.1, there are two primary classes of NORM 583 messages: sender messages and receiver messages. NORM_CMD, NORM_INFO, 584 and NORM_DATA message types are generated by senders of data content, 585 and NORM_NACK and NORM_ACK messages generated by receivers within a 586 NormSession. An auxiliary message type of NORM_REPORT is also provided 587 for experimental purposes. This section describes the message formats 588 used by the NORM protocol. These messages and their fields are 589 referenced in the detailed functional description of the NORM protocol 590 given in Section 5. Individual NORM messages are designed to be 591 compatible with the MTU limitations of encapsulating Internet protocols 592 including IPv4, IPv6, and UDP. The current NORM protocol specification 593 assumes UDP encapsulation and leverages the transport features of UDP. 594 The NORM messages are independent of network addresses and can be used 595 in IPv4 and IPv6 networks. 597 4.1. NORM Common Message Header and Extensions 599 There are some common message fields contained in all NORM message 600 types. Additionally, a header extension mechanism is defined to expand 601 the functionality of the NORM protocol without revision to this 602 document. All NORM protocol messages begin with a common header with 603 information fields as follows: 605 0 1 2 3 606 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 607 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 608 |version| type | hdr_len | sequence | 609 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 610 | source_id | 611 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 613 NORM Common Message Header Format 615 The "version" field is a 4-bit value indicating the protocol version 616 number. NORM implementations SHOULD ignore received messages with 617 version numbers different from their own. This number is intended to 618 indicate and distinguish upgrades of the protocol which may be non- 619 interoperable. The NORM version number for this specification is 1. 621 The message "type" field is a 4-bit value indicating the NORM protocol 622 message type. These types are defined as follows: 624 Message Value 626 NORM_INFO 1 627 NORM_DATA 2 628 NORM_CMD 3 629 NORM_NACK 4 630 NORM_ACK 5 631 NORM_REPORT 6 633 The 8-bit "hdr_len" field indicates the number of 32-bit words that 634 comprise the given message's header portion. This is used to facilitate 635 header extensions that may be applied. The presence of header 636 extensions are implied when the "hdr_len" value is greater than the base 637 value for the given message "type". 639 The "sequence" field is a 16-bit value that is set by the message 640 originator as a monotonically increasing number incremented with each 641 NORM message transmitted to a given destination address. A "sequence" 642 field number space SHOULD be maintained for messages sent to the 643 NormSession group address. This value can be monitored by receiving 644 nodes to detect packet losses in the transmission from a sender and used 645 in estimating raw packet loss for congestion control purposes. Note 646 that this value is NOT used in the NORM protocol to detect missing 647 reliable data content and does NOT identify the application data or FEC 648 that may be attached. With message authentication, the "sequence" field 649 may also be leveraged for protection from message "replay" attacks, 650 particularly of NORM_NACK or other feedback messages. In this case, the 651 receiver node should maintain a monotonically increasing "sequence" 652 field space for each destination to which it transmits (This may be 653 multiple destinations when unicast feedback is used). The size of this 654 field is intended to be sufficient to allow detection of a reasonable 655 range of packet loss within the delay-bandwidth product of expected 656 network connections. 658 The "source_id" field is a 32-bit value identifying the node that sent 659 the message. A participant's NORM node identifier (NormNodeId) can be 660 set according to application needs but unique identifiers must be 661 assigned within a single NormSession. In some cases, use of the host IP 662 address or a hash of it can suffice, but alternative methodologies for 663 assignment and potential collision resolution of node identifiers within 664 a multicast session need to be considered. For example, the "source 665 identifier" mechanism defined in the Real-Time Protocol (RTP) 666 specification [18] may be applicable to use for NORM node identifiers. 667 At this point in time, the protocol makes no assumptions about how these 668 unique identifiers are actually assigned. 670 NORM Header Extensions 672 When header extensions are applied, they follow the message type's base 673 header and precede any payload portion. There are two formats for 674 header extensions, both of which begin with an 8-bit "het" (header 675 extension type) field. One format is provided for variable-length 676 extensions with "het" values in the range from 0 through 127. The other 677 format is for fixed length (one 32-bit word) extensions with "het" 678 values in the range from 128 through 255. These formats are given here: 680 0 1 2 3 681 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 682 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 683 | het <=127 | hel | | 684 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | 685 | Header Extension Content | 686 | ... | 687 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 689 NORM Variable Length Header Extension Format 691 0 1 2 3 692 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 693 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 694 | ext_type >=128| ext_len | Header Extension Content | 695 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 696 NORM Fixed Length (32-bit) Header Extension Format 698 The "Header Extension Content" portion of these header extension format 699 is defined for each header extension type defined for NORM messages. 700 Some header extensions are defined within this document for NORM 701 FEC and congestion control operations. 703 4.2. Sender Messages 705 NORM sender messages include the NORM_DATA type, the NORM_INFO type, and 706 the NORM_CMD type. NORM_DATA and NORM_INFO messages contain application 707 data content while NORM_CMD messages are used for various protocol 708 control functions. 710 4.2.1. NORM_DATA Message 712 The NORM_DATA message is expected to be the predominant type transmitted 713 by NORM senders. These messages are used to encapsulate segmented data 714 content for objects of type NORM_OBJECT_DATA, NORM_OBJECT_FILE, and 715 NORM_OBJECT_STREAM. NORM_DATA messages may contain original or FEC- 716 encoded application data content. 718 The format of NORM_DATA messages is comprised of three logical portions: 719 1) a fixed-format NORM_DATA header portion, 2) a FEC Payload ID portion 720 with a format dependent upon the FEC encoding used, and 3) a payload 721 portion containing source or encoded application data content. Note for 722 objects of type NORM_OBJECT_STREAM, the payload portion contains 723 additional fields used to appropriately recover stream content. NORM 724 implementations MAY also extend the NORM_DATA header to include a FEC 725 Object Transmission Information (EXT_FTI) header extension. This allows 726 NORM receivers to automatically allocate resources and properly perform 727 FEC decoding without the need for pre-configuration or out-of-band 728 information. 730 0 1 2 3 731 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 732 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 733 |version| type=2| hdr_len | sequence | 734 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 735 | source_id | 736 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 737 | instance_id | grtt |backoff| gsize | 738 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 739 | flags | fec_id | object_transport_id | 740 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 741 | fec_payload_id | 742 | ... | 743 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 744 | header_extensions (if applicable) | 745 | ... | 746 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 747 | payload_reserved* | payload_len* | 748 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 749 | payload_offset* | 750 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 751 | payload_data* | 752 | ... | 753 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 755 NORM_DATA Message Format 757 *NOTE: The "payload_reserved", "payload_len" and "payload_offset" 758 fields are present only for objects of type NORM_OBJECT_STREAM. The 759 "payload_len" and "payload_offset" fields allow senders to arbitrarily 760 vary the size of NORM_DATA payload segments for streams. This allows 761 applications to flush transmitted streams as needed to meet unique 762 streaming requirements. For objects of types NORM_OBJECT_FILE and 763 NORM_OBJECT_DATA, these fields are unnecessary since the receiver can 764 calculate the payload length and offset information from the 765 "fec_payload_id" using the algorithm described in Section 5.1.1. The 766 "payload_reserved" field is kept for anticipated future NORM stream 767 control functions. When systematic FEC codes (e.g., "fec_id" = 129) are 768 used, the "payload_len" and "payload_offset" fields contain actual 769 length and offset values for the encapsulated application data segment 770 for those NORM_DATA messages containing source data symbols. In 771 NORM_DATA messages that contain parity information, these fields are not 772 actual length or offset values, but instead are values computed from FEC 773 encoding the "payload_len" and "payload_offset" fields of the _source_ 774 data symbols of the corresponding applicable coding block. 776 The "version", "type", "hdr_len", "sequence", and "source_id" fields 777 form the NORM Common Message Header as described in Section 4.1. The 778 value of the NORM_DATA "type" field is 2. The NORM_DATA _base_ 779 "hdr_len" value is 4 (32-bit words) plus the size of the 780 "fec_payload_id" field. The "fec_payload_id" field size depends upon 781 the FEC encoding used for the referenced NormObject. The "fec_id" field 782 is used to indicate the FEC coding type. For example, when small block, 783 systematic codes are used, a "fec_id" value of 129 is indicated and the 784 of the "fec_payload_id" is two 32-bit words. In this case the NORM_DATA 785 base "hdr_len" value is 6. The cumulative size of any header extensions 786 applied is added into the "hdr_len" field. 788 The "instance_id" field contains a value generated by the sender to 789 uniquely identify its current instance of participation in the 790 NormSession. This allows receivers to detect when senders have perhaps 791 left and rejoined a session in progress. When a sender (identified by 792 its "source_id") is detected to have a new "instance_id", the NORM 793 receivers SHOULD drop their previous state on the sender and begin 794 reception anew. 796 The "grtt" field contains a non-linear quantized representation of the 797 sender's current estimate of group round-trip time (GRTT) (This is also 798 referred to as R_max in [19]). This value is used to control timing of 799 the NACK repair process and other aspects of protocol operation as 800 described in this document. The algorithm for encoding and decoding 801 this field is described in the RMT NORM Building Block document [4]. 803 The "backoff" field value is used by receivers to determine the maximum 804 backoff timer value used in the timer-based NORM NACK feedback 805 suppression. This 4-bit field supports values from 0-15 which is 806 multiplied by the sender GRTT to determine the maximum backoff timeout. 807 The "backoff" field informs the receiver set of the sender's backoff 808 factor parameter "Ksender". Recommended values and their use are 809 described in the NORM receiver NACK procedure description in Section 810 5.3. The "gsize" field contains a representation of the sender's 811 current estimate of group size. This 4-bit field can roughly represent 812 values from ten to 500 million where the most significant bit value of 0 813 or 1 represents a mantissa of 1 or 5, respectively and the three least 814 significant bits incremented by one represent a base 10 exponent (order 815 of magnitude). For examples, a field value of "0x0" represents 1.0e+01 816 (10), a value of "0x8" represents 5.0e+01 (50), a value of "0x1" 817 represents 1.0e+02 (100), and a value of "0xf" represents 5.0e+08. For 818 NORM feedback suppression purposes, the group size does not need to be 819 represented with a high degree of precision. The group size may even be 820 estimated somewhat conservatively (i.e. overestimated) to maintain low 821 levels of feedback traffic. A default group size estimate of 10,000 822 ("gsize" = 0x4) is recommended for general purpose reliable multicast 823 applications using the NORM protocol. 825 The "flags" field contains a number of different binary flags providing 826 information and hints regarding how the receiver should handle the 827 identified object. Defined flags in this field include: 829 +---------------------+-------+------------------------------------------+ 830 | Flag | Value | Purpose | 831 +---------------------+-------+------------------------------------------+ 832 |NORM_FLAG_REPAIR | 0x01 | Indicates message is a repair | 833 | | | transmission | 834 +---------------------+-------+------------------------------------------+ 835 |NORM_FLAG_EXPLICIT | 0x02 | Indicates a repair segment intended to | 836 | | | meet a specific receiver erasure, as | 837 | | | compared to parity segments provided by | 838 | | | the sender for general purpose (with | 839 | | | respect to an FEC coding block) erasure | 840 | | | filling. | 841 +---------------------+-------+------------------------------------------+ 842 |NORM_FLAG_INFO | 0x04 | Indicates availability of NORM_INFO for | 843 | | | object. | 844 +---------------------+-------+------------------------------------------+ 845 |NORM_FLAG_UNRELIABLE | 0x08 | Indicates that repair transmissions for | 846 | | | the specified object will be unavailable | 847 | | | (One-shot, best effort transmission). | 848 +---------------------+-------+------------------------------------------+ 849 |NORM_FLAG_FILE | 0x10 | Indicates object is "file-based" data | 850 | | | (hint to use disk storage for | 851 | | | reception). | 852 +---------------------+-------+------------------------------------------+ 853 |NORM_FLAG_STREAM | 0x20 | Indicates object is of type | 854 | | | NORM_OBJECT_STREAM. | 855 +---------------------+-------+------------------------------------------+ 856 |NORM_FLAG_MSG_START | 0x40 | Marks the first segment of application | 857 | | | messages embedded in | 858 | | | NORM_OBJECT_STREAMs. | 859 +---------------------+-------+------------------------------------------+ 861 NORM_FLAG_REPAIR is set when the associated message is a repair 862 transmission. This information can be used by receivers to help observe 863 a join policy where it is desired that newly joining receivers only 864 begin participating in the NACK process upon receipt of new (non-repair) 865 data content. NORM_FLAG_EXPLICIT is used to mark repair messages sent 866 when the data sender has exhausted its ability to provide "fresh" 867 (previously untransmitted) parity segments as repair. This flag could 868 possibly be used by intermediate systems implementing functionality to 869 control sub-casting of repair content to different legs of a reliable 870 multicast topology with disparate repair needs. NORM_FLAG_INFO is set 871 only when optional NORM_INFO content is actually available for the 872 associated object. Thus, receivers will NACK for retransmission of 873 NORM_INFO only when it is available for a given object. 874 NORM_FLAG_UNRELIABLE is set when the sender wishes to transmit an object 875 with only "best effort" delivery and will not supply repair 876 transmissions for the object. NORM receivers SHOULD NOT execute repair 877 requests for objects marked with the NORM_FLAG_UNRELIABLE flag. Note 878 receivers may inadvertently request repair of such objects when all 879 segments (or info content) for those objects are not received (i.e. a 880 gap in the "object_transport_id" sequence is noted). In this case, the 881 sender should invoke the NORM_CMD(SQUELCH) process as described in 882 Section 4.2.3. NORM_FLAG_FILE can be set as a "hint" from the sender 883 that the associated object should be stored in non-volatile storage. 884 NORM_FLAG_STREAM is set when the identified object is of type 885 NORM_OBJECT_STREAM. When NORM_FLAG_STREAM is set, the 886 NORM_FLAG_MSG_START can be optionally used to mark the first data 887 segments of application-layer messages transported within the NORM 888 stream. This allows NORM receiver applications to "synchronize" with 889 NORM senders and to be able to properly interpret application layer data 890 when joining a NORM session already in progress. In practice, the NORM 891 implementation MAY set this flag for the segment transmitted following 892 an explicit "flush" of the stream by the application. 894 The "fec_id" field corresponds to the FEC Encoding Identifier described 895 in the FEC Building Block document [5]. The "fec_id" value implies the 896 format of the "fec_payload_id" field and, coupled with FEC Object 897 Transmission Information, the procedures to decode FEC encoded content. 898 Small block, systematic codes ("fec_id" = 129) are expected to be used 899 for most NORM purposes and the NORM_OBJECT_STREAM requires systematic 900 FEC codes for most efficient performance. 902 The "object_transport_id" field is a monotonically and incrementally 903 increasing value assigned by the sender to NormObjects being 904 transmitted. Transmissions and repair requests related to that object 905 use the same "object_transport_id" value. For sessions of very long or 906 indefinite duration, the "object_transport_id" field may be repeated, 907 but it is presumed that the 16-bit field size provides an adequate 908 enough sequence space to avoid object confusion amongst receivers and 909 sources (i.e. receivers SHOULD re-synchronize with a server when 910 receiving object sequence identifiers sufficiently out-of-range with the 911 current state kept for a given source). During the course of its 912 transmission within a NORM session, an object is uniquely identified by 913 the concatenation of the sender "source_id" and the given 914 "object_transport_id". Note that NORM_INFO messages associated with the 915 identified object carry the same "object_transport_id" value. 917 The "fec_payload_id" identifies the attached NORM_DATA "payload" 918 content. The size and format of the "fec_payload_id" field depends upon 919 the FEC type indicated by the "fec_id" field. These formats are given 920 in the FEC Building Block document [5] and any subsequent extensions of 921 that document. As an example, the format of the "fec_payload_id" format 922 small block, systematic codes ("fec_id" = 129) given here: 924 0 1 2 3 925 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 926 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 927 | source_block_number | 928 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 929 | source_block_len | encoding_symbol_id | 930 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 932 Small Block, Systematic Code ("fec_id" = 129) "fec_payload_id" Format 934 The FEC payload identifier "source_block_number", "source_block_len", 935 and "encoding_symbol_id" fields correspond to the "Source Block Number", 936 "Source Block Length, and "Encoding Symbol ID" fields of the FEC Payload 937 ID format given by the IETF FEC Building Block document [5]. The 938 "source_block_number" identifies the coding block's relative position 939 with a NormObject. Note that, for NormObjects of type 940 NORM_OBJECT_STREAM, the "source_block_number" may wrap for very long 941 lived sessions. The "source_block_len" indicates the number of user 942 data segments in the identified coding block. Given the 943 "source_block_len" information of how many symbols of application data 944 are contained in the block, the receiver can determine whether the 945 attached segment is data or parity content and treat it appropriately. 946 The "encoding_symbol_id" identifies which specific symbol (segment) 947 within the coding block the attached payload conveys. Depending upon 948 the value of the "encoding_symbol_id" and the associated 949 "source_block_len" parameters for the block, the symbol (segment) 950 referenced may be a user data or an FEC parity segment. For systematic 951 codes, encoding symbols numbered less than the source_block_len contain 952 original application data while segments greater than or equal to 953 source_block_len contain parity symbols calculated for the block. The 954 concatenation of object_transport_id::fec_payload_id can be viewed as a 955 unique transport protocol data unit identifier for the attached segment 956 with respect to the NORM sender's instance within a session. 958 Additional FEC Object Transmission Information (as described in the FEC 959 Building Block document [5]) is required to properly receive and decode 960 NORM transport objects. This information MAY be provided as out-of-band 961 session information. However, in some cases, it may be useful for the 962 sender to include this information "in band" to facilitate receiver 963 operation with minimal preconfiguration. For this purpose, the NORM FEC 964 Object Transmission Information Header Extension (EXT_FTI) is defined. 965 This header extension MAY be applied to NORM_DATA and NORM_INFO messages 966 to provide this necessary information. The exact format of the 967 extension depends upon the FEC code in use, but in general it SHOULD 968 contain any required details on the FEC code in use (e.g., FEC Instance 969 ID, etc) and the byte size of the associated NormObject (For the 970 NORM_OBJECT_STREAM type, this size corresponds to the stream buffer size 971 maintained by the NORM sender). As an example, the format of the 972 EXT_FTI for small block systematic codes ("fec_id" = 129) is given here: 974 0 1 2 3 975 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 976 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 977 | het = 64 | hel = 4 | object_length (msb) | 978 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 979 | object_length (lsb) | 980 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 981 | fec_instance_id | segment_size | 982 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 983 | fec_max_block_len | fec_num_parity | 984 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 986 FEC Object Transmission Information Header Extension (EXT_FTI) for Small Block Systematic Codes ("fec_id" = 129) 988 The header extension type "het" field value for this header extension is 989 64. The header extension length "hel" depends upon the format of the 990 FTI for FEC code type identified by the "fec_id" field. In this example 991 (for "fec_id" = 129), the "hel" field value is 4. 993 The 48-bit "object_length" field indicates the total size of the object 994 (in bytes) for the static object types of NORM_OBJECT_FILE and 995 NORM_OBJECT_DATA. This information is used by receivers to determine 996 storage requirements and/or allocate storage for the received object. 997 Receivers with insufficient storage capability may wish to forego 998 reliable reception (i.e. not NACK for) of the indicated object. In the 999 case of objects of type NORM_OBJECT_STREAM, the "object_length" field is 1000 used by the sender to indicate the size of its stream buffer to the 1001 receiver group. In turn, the receivers SHOULD use this information to 1002 allocate a stream buffer for reception of corresponding size. 1004 The "fec_instance_id" corresponds to the "FEC Instance ID" described in 1005 the FEC Building Block document [5]. In this case, the 1006 "fec_instance_id" SHALL be a value corresponding to the particular type 1007 of Small Block Systematic Code being used (e.g., Reed-Solomon GF(2^8), 1008 Reed-Solomon GF(2^16), etc). The standardized assignment of FEC 1009 Instance ID values is described in [5]. The "segment_size" field 1010 indicates the sender's current setting for maximum message payload 1011 content (in bytes). This allows receivers to allocate appropriate 1012 buffering resources and to determine other information in order to 1013 properly process received data messaging. 1015 The "fec_max_block_len" indicates the current maximum number of user 1016 data segments per FEC coding block to be used by the sender during the 1017 session. This allows receivers to allocate appropriate buffer space for 1018 buffering blocks transmitted by the sender. 1020 The "fec_num_parity" corresponds to the "maximum number of encoding 1021 symbols that can be generated for any source block" as described in for 1022 FEC Object Transmission Information for Small Block Systematic Codes in 1023 the FEC Building Block document [5]. For example, Reed-Solomon codes 1024 may be arbitrarily shortened to create different code variations for a 1025 given block length. In the case of Reed-Solomon (GF(2^8) and GF(2^16) 1026 codes, this value indicates the maximum number of parity segments 1027 available from the sender for the coding blocks. This field MAY be 1028 differently for other systematic codes as they are defined. 1030 The payload portion of NORM_DATA messages includes source data or FEC 1031 encoded application content. 1033 The "payload_reserved", "payload_len" and "payload_offset" fields are 1034 present ONLY for transport objects of type NORM_OBJECT_STREAM. These 1035 fields indicated the size and relative position (within the stream) of 1036 the application content represented by the message payload. For senders 1037 employing systematic FEC encoding, these fields contain _actual_ length 1038 and offset values (in bytes) for the payload of messages which contain 1039 original data source symbols. For NORM_DATA messages containing 1040 calculated parity content, these fields will actually contain values 1041 computed by FEC encoding of the "payload_len" and "payload_offset" 1042 values of the NORM_DATA data segments of the corresponding FEC coding 1043 block. Thus, the "payload_len" and "payload_offset" values of missing 1044 data content can be determined upon decoding a FEC coding block. Note 1045 that these fields do NOT contribute to the value of the NORM_DATA 1046 "hdr_len" field. These fields are NOT present when the "flags" portion 1047 of the NORM_DATA| message indicate the transport object if of type 1048 NORM_OBJECT_FILE or NORM_OBJECT_DATA. In this case, the length and 1049 offset information can be calculated from the "fec_payload_id" using the 1050 methodology described in Section 5.1.1. Note that for long-lived 1051 streams, the "payload_offset" field can wrap. 1053 The "payload_data" field contains the original application source or 1054 parity content for the symbol identified by the "fec_payload_id". The 1055 length of this field SHALL be limited to a maximum of the sender's 1056 NormSegmentSize bytes as given in the FTI for the object. Note the 1057 length of this field for messages containing parity content will always 1058 be of length NormSegmentSize. When encoding data segments of varying 1059 sizes, the FEC encoder SHALL assume ZERO value padding for data segments 1060 with length less than the NormSegmentSize. It is RECOMMENDED that a 1061 sender's NormSegmentSize generally be constant for the duration of a 1062 given sender's term of participation in the session, but may possibly 1063 vary on a per-object basis. The NormSegmentSize is expected to be 1064 configurable by the sender application prior to session participation as 1065 needed for network topology maximum transmission unit (MTU) 1066 considerations. For IPv6, MTU discovery may be possibly leveraged at 1067 session startup to perform this configuration. The "payload_data" 1068 content may be delivered directly to the application for source symbols 1069 (when systematic FEC encoding is used) or upon decoding of the FEC 1070 block. For NORM_OBJECT_FILE and NORM_OBJECT_STREAM objects, the data 1071 segment length and offset can be calculated using the algorithm 1072 described in Section 5.1.1. For NORM_OBJECT_STREAM objects, the length 1073 and offset is obtained from the segment's corresponding "payload_len" 1074 and "payload_offset" fields. 1076 4.2.2. NORM_INFO Message 1078 The NORM_INFO message is used to convey OPTIONAL, application-defined, 1079 "out-of-band" context information for transmitted NormObjects. An 1080 example NORM_INFO use for bulk file transfer is to place MIME type 1081 for the associated file, data, or stream object into the NORM_INFO 1082 payload. Receivers may use the NORM_INFO content to make a decision as 1083 whether to participate in reliable reception of the associated object. 1084 Each NormObject can have an independent unit of NORM_INFO associated 1085 with it. NORM_DATA messages contain a flag to indicate the availability 1086 of NORM_INFO for a given NormObject. NORM receivers may NACK for 1087 retransmission of NORM_INFO when they have not received it for a given 1088 NormObject. The size of the NORM_INFO content is limited to that of a 1089 single NormSegmentSize for the given sender. This atomic nature allows 1090 the NORM_INFO to be rapidly and efficiently repaired within the NORM 1091 reliable transmission process. 1093 When NORM_INFO content is available for a NormObject, the NORM_FLAG_INFO 1094 flag SHALL be set in NORM_DATA messages for the corresponding 1095 "object_transport_id" and the NORM_INFO message shall be transmitted as 1096 the first message for the NormObject. 1098 0 1 2 3 1099 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 1100 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1101 |version| type=1| hdr_len | sequence | 1102 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1103 | source_id | 1104 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1105 | instance_id | grtt |backoff| gsize | 1106 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1107 | flags | fec_id | object_transport_id | 1108 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1109 | header_extensions (if applicable) | 1110 | ... | 1111 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1112 | payload_data | 1113 | ... | 1114 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1116 NORM_INFO Message Format 1118 The "version", "type", "hdr_len", "sequence", and "source_id" fields 1119 form the NORM Common Message Header as described in Section 4.1. The 1120 value of "hdr_len" field when no header extensions are present is 4. 1122 The "instance_id", "grtt", "backoff", "gsize", "flags", "fec_id", and 1123 "object_transport_id" fields carry the same information and serve the 1124 same purpose as with NORM_DATA messages. These values allow the 1125 receiver to prepare appropriate buffering, etc, for further 1126 transmissions from the sender when NORM_INFO is the first message 1127 received. 1129 As with NORM_DATA messages, the NORM FTI Header Extension (EXT_FTI) may 1130 be optionally applied to NORM_INFO messages. To conserve protocol 1131 overhead, some NORM implementations may wish to apply the EXT_FTI when 1132 used to NORM_INFO messages only and not to NORM_DATA messages. 1134 The NORM_INFO "payload_data" field contains sender application-defined 1135 content which can be used by receiver applications for various purposes 1136 as described above. 1138 4.2.3. NORM_CMD Messages 1140 NORM_CMD messages are transmitted by senders to perform a number of 1141 different protocol functions. This includes functions such as round- 1142 trip timing collection, congestion control functions, synchronization of 1143 sender/receiver repair "windows", and notification of sender status. A 1144 core set of NORM_CMD messages is enumerated. Additionally, a range of 1145 command types remain available for potential application-specific use. 1146 Some NORM_CMD types may have dynamic content attached. Any attached 1147 content will be limited to maximum length of the sender NormSegmentSize 1148 to retain the atomic nature of commands. All NORM_CMD messages begin 1149 with a common set of fields, after the usual NORM message common header. 1150 The standard NORM_CMD fields are: 1152 0 1 2 3 1153 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 1154 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1155 |version| type=3| hdr_len | sequence | 1156 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1157 | source_id | 1158 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1159 | instance_id | grtt |backoff| gsize | 1160 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1161 | flavor | | 1162 +-+-+-+-+-+-+-+-+ NORM_CMD Content + 1163 | ... | 1164 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1166 NORM_CMD Standard Fields 1168 The "version", "type", "hdr_len", "sequence", and "source_id" fields 1169 form the NORM Common Message Header as described in Section 4.1. The 1170 value of the "hdr_len" field for NORM_CMD messages without header 1171 extensions present depends upon the "flavor" field. 1173 The "instance_id", "grtt", "backoff", and "gsize" fields provide the 1174 same information and serve the same purpose as with NORM_DATA and 1175 NORM_INFO messages. The "flavor" field indicates the type of command to 1176 follow. The remainder of the NORM_CMD message is dependent upon the 1177 command type ("flavor"). NORM command flavors include: 1179 +----------------------+--------------+----------------------------------+ 1180 | Command | Flavor Value | Purpose | 1181 +----------------------+--------------+----------------------------------+ 1182 |NORM_CMD(FLUSH) | 1 | Used to indicate sender | 1183 | | | temporary end-of-transmission. | 1184 | | | (Assists in robustly initiating | 1185 | | | outstanding repair requests from | 1186 | | | receivers). May also be | 1187 | | | optionally used to collect | 1188 | | | positive acknowledgment of | 1189 | | | reliable reception from subset | 1190 | | | of receivers. | 1191 +----------------------+--------------+----------------------------------+ 1192 |NORM_CMD(EOT) | 2 | Used to indicate sender | 1193 | | | permanent end-of-transmission. | 1194 +----------------------+--------------+----------------------------------+ 1195 |NORM_CMD(SQUELCH) | 3 | Used to advertise sender's | 1196 | | | current repair window in | 1197 | | | response to out-of-range NACKs | 1198 | | | from receivers. | 1199 +----------------------+--------------+----------------------------------+ 1200 |NORM_CMD(CC) | 4 | Used for GRTT measurement and | 1201 | | | collection of congestion control | 1202 | | | feedback. | 1203 +----------------------+--------------+----------------------------------+ 1204 |NORM_CMD(REPAIR_ADV) | 5 | Used to advertise sender's | 1205 | | | aggregated repair/feedback state | 1206 | | | for suppression of unicast | 1207 | | | feedback from receivers. | 1208 +----------------------+--------------+----------------------------------+ 1209 |NORM_CMD(ACK_REQ) | 6 | Used to request application- | 1210 | | | defined positive acknowledgment | 1211 | | | from a list of receivers | 1212 | | | (OPTIONAL). | 1213 +----------------------+--------------+----------------------------------+ 1214 |NORM_CMD(APPLICATION) | 7 | Used for application-defined | 1215 | | | purposes which may need to | 1216 | | | temporarily preempt data | 1217 | | | transmission (OPTIONAL). | 1218 +----------------------+--------------+----------------------------------+ 1220 4.2.3.1. NORM_CMD(FLUSH) Message 1222 The NORM_CMD(FLUSH) command is sent when the sender reaches the end of 1223 all data content and pending repairs it has queued for transmission. 1224 This may indicate a temporary or permanent end of data transmission, but 1225 the sender is still willing to respond to repair requests. This command 1226 is repeated once per 2*GRTT to excite the receiver set for any 1227 outstanding repair requests up to and including the transmission point 1228 indicated within the NORM_CMD(FLUSH) message. The number of repeats is 1229 equal to NORM_ROBUST_FACTOR unless a list of receivers from which 1230 explicit positive acknowledgment ("acking_node_list") is given. In that 1231 case, the "acking_node_list" is updated as acknowledgments are received 1232 the NORM_CMD(FLUSH) is repeated according to the mechanism described in 1233 Section 5.5.3. The greater the NORM_ROBUST_FACTOR, the greater the 1234 probability that all applicable receivers will be excited for 1235 acknowledgment or repair requests (NACKs) _and_ that the corresponding 1236 NACKs are delivered to the sender. If a NORM_NACK message interrupts 1237 the flush process, the sender will re-initiate the flush process after 1238 any resulting repair transmissions are completed. 1240 Note that receivers also employ a timeout mechanism to self-initiate 1241 NACKing (if there are outstanding repair needs) when no messages of any 1242 type are received from a sender. This inactivity timeout is related to 1243 2*GRTT*NORM_ROBUST_FACTOR and will be discussed more later. With a 1244 sufficient NORM_ROBUST_FACTOR value, data content is delivered with a 1245 high assurance of reliability. The penalty of a large 1246 NORM_ROBUST_FACTOR value is potentially excess sender NORM_CMD(FLUSH) 1247 transmissions and a longer timeout for receivers to self-initiate the 1248 terminal NACK process. 1250 For finite-size transport objects such as NORM_OBJECT_DATA and 1251 NORM_OBJECT_FILE, the flush process (if there are no further pending 1252 objects) occurs at the end of these objects. Thus, FEC repair 1253 information is always available for repairs in response to repair 1254 requests elicited by the flush command. However, for 1255 NORM_OBJECT_STREAM, the flush may occur at any time, including in the 1256 middle of an FEC coding block if systematic FEC codes are employed. In 1257 this case, the sender will not yet be able to provide FEC parity content 1258 as repair for the concurrent coding block and will be limited to 1259 explicitly repairing stream data content for that block. Applications 1260 that anticipate frequent flushing of stream content SHOULD be judicious 1261 in the selection of the FEC coding block size (i.e., do not use a very 1262 large coding block size if frequent flushing occurs). For example, a 1263 reliable multicast application transmitting an on-going series of 1264 intermittent, relatively small messaging content will need to trade-off 1265 using the NORM_OBJECT_DATA paradigm versus the NORM_OBJECT_STREAM 1266 paradigm with an appropriate FEC coding block size. This is analogous 1267 to application trade-offs for other transport protocols such as the 1268 selection of different TCP modes of operation such as "no delay", etc. 1270 0 1 2 3 1271 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 1272 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1273 |version| type=3| hdr_len | sequence | 1274 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1275 | source_id | 1276 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1277 | instance_id | grtt |backoff| gsize | 1278 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1279 | flavor = 1 | fec_id | object_transport_id | 1280 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1281 | fec_payload_id | 1282 | ... | 1283 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1284 | acking_node_list (if applicable) | 1285 | ... | 1286 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1288 NORM_CMD(FLUSH) Message Format 1290 In addition to the NORM common message header and standard NORM_CMD 1291 fields, the NORM_CMD(FLUSH) message contains fields to identify the 1292 current status and logical transmit position of the sender. 1294 The "fec_id" field indicates the FEC type used for the flushing 1295 "object_transport_id" and implies the size and format of the 1296 "fec_payload_id" field. Note the "hdr_len" value for the 1297 NORM_CMD(FLUSH) message is 4 plus the size of the "fec_payload_id" field 1298 when no header extensions are present. 1300 The "object_transport_id" and "fec_payload_id" fields indicate the 1301 sender's current logical "transmit position". These fields are 1302 interpreted in the same manner as in the NORM_DATA message type. Upon 1303 receipt of the NORM_CMD(FLUSH), receivers are expected to check their 1304 completion state _through_ (including) this transmission position. If 1305 receivers have outstanding repair needs in this range, they SHALL 1306 initiate the NORM NACK Repair Process as described in Section 5.3. If 1307 receivers have no outstanding repair needs, no response to the 1308 NORM_CMD(FLUSH) is generated. 1310 For NORM_OBJECT_STREAM objects using systematic FEC codes, receivers 1311 MUST request "explicit-only" repair of the identified 1312 "source_block_number" if the given "encoding_symbol_id" is less than the 1313 "source_block_len". This condition indicates the sender has not yet 1314 completed encoding the corresponding FEC block and parity content is not 1315 yet available. An "explicit-only" repair request consists of NACK 1316 content for the applicable "source_block_number" which does not include 1317 any requests for parity-based repair. This allows NORM sender 1318 applications to "flush" an ongoing stream of transmission when needed, 1319 even if in the middle of an FEC block. Once the sender resumes stream 1320 transmission and passes the end of the pending coding block, subsequent 1321 NACKs from receivers SHALL request parity-based repair as usual. Note 1322 that the use of a systematic FEC code is assumed here. Normal receiver 1323 NACK initiation and construction is discussed in detail in Section 5.3. 1325 OPTIONAL "acking_node_list" field contains a list of NormNodeIds for 1326 receivers from which the sender is requesting explicit positive 1327 acknowledgment of reception up through the transmission point identified 1328 by the "object_transport_id" and "fec_payload_id" fields. The length of 1329 the list can be inferred from the length of the received NORM_CMD(FLUSH) 1330 message. When the "acking_node_list" is present, the lightweight 1331 positive acknowledgment process described in Section 5.5.3 SHALL be 1332 observed. 1334 4.2.3.2. NORM_CMD(EOT) Message 1336 The NORM_CMD(EOT) command is sent when the sender reaches permanent end- 1337 of-transmission with respect to the NormSession and will not respond to 1338 further repair requests. This allows receivers to gracefully reach 1339 closure of operation with this sender (without requiring any timeout) 1340 and free any resources that are no longer needed. The NORM_CMD(EOT) 1341 command SHOULD be sent with the same robust mechanism as used for 1342 NORM_CMD(FLUSH) commands to provide a high assurance of reception by the 1343 receiver set. 1345 0 1 2 3 1346 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 1347 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1348 |version| type=3| hdr_len | sequence | 1349 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1350 | source_id | 1351 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1352 | instance_id | grtt |backoff| gsize | 1353 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1354 | flavor = 2 | reserved | 1355 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1357 NORM_CMD(EOT) Message Format 1359 The value of the "hdr_len" field for NORM_CMD(EOT) messages without 1360 header extensions present is 4. The "reserved" field is reserved for 1361 future use and MUST be set to an all ZERO value. Receivers MUST ignore 1362 the "reserved" field. 1364 4.2.3.3. NORM_CMD(SQUELCH) Message 1366 The NORM_CMD(SQUELCH) command is transmitted in response to outdated or 1367 invalid NORM_NACK content received by the sender. Invalid NORM_NACK 1368 content consists of repair requests for NormObjects for which the sender 1369 is unable or unwilling to provide repair. This includes repair requests 1370 for outdated objects, aborted objects, or those objects which the sender 1371 previously transmitted marked with the NORM_FLAG_UNRELIABLE flag. This 1372 command indicates to receivers what content is available for repair, 1373 thus serving as a description of the sender's current "repair window". 1374 Receivers SHALL not generate repair requests for content identified as 1375 invalid by a NORM_CMD(SQUELCH). 1377 The NORM_CMD(SQUELCH) command is sent once per 2*GRTT at the most. The 1378 NORM_CMD(SQUELCH) advertises the current "repair window" of the sender 1379 by identifying the earliest (lowest) transmission point for which it 1380 will provide repair, along with an encoded list of objects from that 1381 point forward that are no longer valid for repair. This mechanism 1382 allows the sender application to cancel or abort transmission and/or 1383 repair of specific previously enqueued objects. The list also contains 1384 the identifiers for any objects within the repair window that were sent 1385 with the NORM_FLAG_UNRELIABLE flag set. In normal conditions, it is 1386 expected the NORM_CMD(SQUELCH) will be needed infrequently, and 1387 generally only to provide a reference repair window for receivers who 1388 have fallen "out-of-sync" with the sender due to extremely poor network 1389 conditions. 1391 The starting point of the invalid NormObject list begins with the lowest 1392 invalid NormTransportId greater than the current "repair window" start 1393 from the invalid NACK(s) that prompted the generation of the squelch. 1394 The length of the list is limited by the sender's NormSegmentSize. This 1395 allows the receivers to learn the status of the sender's applicable 1396 object repair window with minimal transmission of NORM_CMD(SQUELCH) 1397 commands. The format of the NORM_CMD(SQUELCH) message is: 1399 0 1 2 3 1400 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 1401 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1402 | version | type = 3 | sequence | 1403 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1404 | source_id | 1405 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1406 | instance_id | grtt |backoff| gsize | 1407 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1408 | flavor = 3 | fec_id | object_transport_id | 1409 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1410 | fec_payload_id | 1411 | ... | 1412 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1413 | invalid_object_list | 1414 | ... | 1415 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1417 NORM_CMD(SQUELCH) Message Format 1419 In addition to the NORM common message header and standard NORM_CMD 1420 fields, the NORM_CMD(SQUELCH) message contains fields to identify the 1421 earliest logical transmit position of the sender's current repair window 1422 and an "invalid object list" beginning with the index of the logically 1423 earliest invalid repair request from the offending NACK message which 1424 initiated the squelch transmission. 1426 The "object_transport_id" and "fec_payload_id" fields are concatenated 1427 to indicate the beginning of the sender's current repair window (i.e., 1428 the logically earliest point in its transmission history for which the 1429 sender can provide repair). The "fec_id" field implies the size and 1430 format of the "fec_payload_id" field. This serves as an advertisement 1431 a "synchronization point" for receivers to request repair. Note, that 1432 while an "encoding_symbol_id" may be included in the "fec_payload_id" 1433 field, the sender's repair window SHOULD be aligned on FEC coding block 1434 boundaries and thus the "encoding_symbol_id" SHOULD be zero. 1436 The "invalid_object_list" is a list of 16-bit NormTransportIds that, 1437 although they are within the range of the sender's current repair 1438 window, are no longer available for repair from the sender. For example, 1439 a sender application may dequeue an out-of-date object even though it is 1440 still within the repair window. The total size of the 1441 "invalid_object_list" content is can be determined from the packet's 1442 payload length and is limited to a maximum of the NormSegmentSize of the 1443 sender. Thus, for very large repair windows, it is possible that a 1444 single NORM_CMD(SQUELCH) message may not be capable of listing the 1445 entire set of invalid objects in the repair window. In this case, the 1446 sender SHALL ensure that the list begins with a NormObjectId that is 1447 greater than or equal to the lowest ordinal invalid NormObjectId from 1448 the NACK message(s) that prompted the NORM_CMD(SQUELCH) generation. The 1449 NormObjectIds in the "invalid_object_list" MUST be greater than the 1450 "object_transport_id" marking the beginning of the sender's repair 1451 window. This insures convergence of the squelch process, even if 1452 multiple invalid NACK/ squelch iterations are required. This explicit 1453 description of invalid content within the sender's current window allows 1454 the sender application (most notably for discrete "object" based 1455 transport) to arbitrarily invalidate (i.e. dequeue) portions of enqueued 1456 content (e.g., certain objects) for which it no longer wishes to provide 1457 reliable transport. 1459 4.2.3.4. NORM_CMD(CC) Message 1461 The NORM_CMD(CC) messages contains fields to enable sender-to-receiver 1462 group greatest round-trip time (GRTT) measurement and to excite the 1463 group for congestion control feedback. A baseline NORM congestion 1464 control scheme (NORM-CC), based on the TCP-Friendly Multicast Congestion 1465 Control (TFMCC) scheme of [19] is described in Section 5.5.2 of this 1466 document. The NORM_CMD(CC) message is usually transmitted as part of 1467 NORM-CC congestion control operation. A NORM header extension is 1468 defined below to be used with the NORM_CMD(CC) message to support NORM- 1469 CC operation. Different header extensions may be defined for the 1470 NORM_CMD(CC) (and/or other NORM messages as needed) to support 1471 alternative congestion control schemes in the future. If NORM is 1472 operated in a private network with congestion control operation 1473 disabled, the NORM_CMD(CC) message is then used for GRTT measurement 1474 only and may optionally be sent less frequently than with congestion 1475 control operation. 1477 0 1 2 3 1478 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 1479 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1480 |version| type=3| hdr_len | sequence | 1481 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1482 | source_id | 1483 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1484 | instance_id | grtt |backoff| gsize | 1485 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1486 | flavor = 4 | reserved | cc_sequence | 1487 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1488 | send_time_sec | 1489 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1490 | send_time_usec | 1491 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1492 | header extensions (if applicable) | 1493 | ... | 1494 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1495 | cc_node_list (if applicable) | 1496 | ... | 1497 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1499 NORM_CMD(CC) Message Format 1501 The NORM common message header and standard NORM_CMD fields serve their 1502 usual purposes. 1504 The "reserved" field is for potential future use and should be set to 1505 ZERO in this version of the NORM protocol. 1507 The "cc_sequence" field is a sequence number applied by the sender. For 1508 NORM-CC operation, it is used to provide functionality equivalent to the 1509 "feedback round number" (fb_nr)described in [19]. The most recently 1510 received "cc_sequence" value is recorded by receivers and can be fed 1511 back to the sender in congestion control feedback generated by the 1512 receivers for that sender. The "cc_sequence" number can also be used in 1513 NORM implementations to assess how recently a receiver has received 1514 NORM_CMD(CC) probes from the sender. This can be useful instrumentation 1515 for complex or experimental multicast routing environments. 1517 The "send_time" field is a timestamp indicating the time that the 1518 NORM_CMD(CC) message was transmitted. This consists of a 64-bit field 1519 containing 32-bits with the time in seconds ("send_time_sec") and 1520 32-bits with the time in microseconds ("send_time_usec") since some 1521 reference time the source maintains (usually 00:00:00, 1 January 1970). 1522 The byte ordering of the fields is "Big Endian" network order. 1523 Receivers use this timestamp adjusted by the amount of delay from the 1524 time they received the NORM_CMD(CC) message to the time of their 1525 response as the "grtt_response" portion of NORM_ACK and NORM_NACK 1526 messages generated. This allows the sender to evaluate round-trip times 1527 to different receivers for congestion control and other (e.g., GRTT 1528 determination) purposes. 1530 To facilitate the baseline NORM-CC scheme described in Section 5.5.2, a 1531 Rate header extension (EXT_RATE) is defined to inform the group of the 1532 sender's current transmission rate. This is used along with the loss 1533 detection "sequence" field of all NORM sender messages and the 1534 NORM_CMD(CC) GRTT collection process to support NORM-CC congestion 1535 control operation. The format of this header extension is as follows: 1537 0 1 2 3 1538 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 1539 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1540 | ext_type = 128| reserved | send_rate | 1541 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1543 NORM-CC Rate Header Extension Format (EXT_RATE) 1545 The "send_rate" field indicates the sender's current transmission rate 1546 in bytes per second. The 16-bit "send_rate" field consists of 12 bits 1547 of mantissa in the most significant portion and 4 bits of base 10 1548 exponent (order of magnitude) information in the least significant 1549 portion. The 12-bit mantissa portion of the field is scaled such that a 1550 floating point value of 0.0 corresponds to 0 and a floating point value 1551 of 10.0 corresponds to 4096. Thus: 1553 send_rate = (((int)(Value_mantissa * 4096.0 / 10.0 + 0.5)) << 4) | Value_exponent; 1555 For example, to represent a transmission rate of 256kbps (3.2e+04 bytes 1556 per second), the lower 4 bits of the 16-bit field contain a value of 1557 0x04 to represent the exponent while the upper 12 bits contain a value 1558 of 0x51f as determined from the equation given above: 1560 send_rate = (((int)((3.2 * 4096.0 / 10.0) + 0.5)) << 4) | 4; 1562 = (0x51f << 4) | 0x4 1564 = 0x51f4 1566 To decode the "send_rate" field, the following equation can be used: 1568 value = (send_rate >> 4) * 10.0 / 4096.0 * power(10.0, (send_rate & x000f)) 1570 Note the maximum transmission rate that can be represented by this 1571 scheme is approximately 9.99e+15 bytes per second. 1573 When this extension is present, a "cc_node_list" may be attached as the 1574 payload of the NORM_CMD(CC) message. The presence of this header 1575 extension also implies that NORM receivers should respond according to 1576 the procedures described in Section 5.5.2. The "cc_node_list" consists 1577 of a list of NormNodeIds and their associated congestion control status. 1578 This includes the current limiting receiver (CLR) node, any potential 1579 limiting receiver (PLR) nodes that have been identified, and some number 1580 of receivers for which congestion control status is being provided, most 1581 notably including the receivers' current RTT measurement. The maximum 1582 length of the "cc_node_list" provides for at least the CLR and one other 1583 receiver, but may be configurable for more timely feedback to the group. 1585 list length can be inferred from the length of the NORM_CMD(CC) message. 1587 Each item in the "cc_node_list" is in the following format: 1589 0 1 2 3 1590 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 1591 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1592 | cc_node_id | 1593 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1594 | cc_flags | cc_rtt | cc_rate | 1595 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1597 Congestion Control Node List Item Format 1599 The "cc_node_id" is the NormNodeId of the receiver which the item 1600 represents. 1602 The "cc_flags" field contains flags indicating the congestion control 1603 status of the indicated receiver. The following flags are defined: 1605 +-------------------+-------+------------------------------------------+ 1606 | Flag | Value | Purpose | 1607 +-------------------+-------+------------------------------------------+ 1608 |NORM_FLAG_CC_CLR | 0x01 | Receiver is the current limiting | 1609 | | | receiver (CLR). | 1610 +-------------------+-------+------------------------------------------+ 1611 |NORM_FLAG_CC_PLR | 0x02 | Receiver is a potential limiting | 1612 | | | receiver (PLR). | 1613 +-------------------+-------+------------------------------------------+ 1614 |NORM_FLAG_CC_RTT | 0x04 | Receiver has measured RTT with respect | 1615 | | | to sender. | 1616 +-------------------+-------+------------------------------------------+ 1617 |NORM_FLAG_CC_START | 0x08 | Sender/receiver is in "slow start" phase | 1618 | | | of congestion control operation (i.e. | 1619 | | | The receiver has not yet detected any | 1620 | | | packet loss and the "cc_rate" field is | 1621 | | | the receiver's actual measured receive | 1622 | | | rate). | 1623 +-------------------+-------+------------------------------------------+ 1624 |NORM_FLAG_CC_LEAVE | 0x10 | Receiver is imminently leaving the | 1625 | | | session and its feedback should not be | 1626 | | | considered in congestion control | 1627 | | | operation. | 1628 +-------------------+-------+------------------------------------------+ 1630 The "cc_rtt" contains a quantized representation of the RTT as measured 1631 by the sender with respect to the indicated receiver. This field is 1632 valid only if the NORM_FLAG_CC_RTT flag is set in the "cc_flags" field. 1633 This one byte field is a quantized representation of the RTT using the 1634 algorithm described in the NORM Building Block document [4]. The 1635 "cc_rate" field contains a representation of the receiver's current 1636 calculated (during steady-state congestion control operation) or twice 1637 measured (during the "slow start" phase) congestion control rate. This 1638 field is encoded and decoded using the same technique as described for 1639 the NORM_CMD(CC) "send_rate" field. 1641 4.2.3.5. NORM_CMD(REPAIR_ADV) Message 1643 The NORM_CMD(REPAIR_ADV) message is used by the sender to "advertise" 1644 its aggregated repair state from NORM_NACK messages accumulated during a 1645 repair cycle and/or congestion control feedback received. This message 1646 is sent only when the sender has received NORM_NACK and/or NORM_ACK(CC) 1647 (when congestion control is enabled) messages via unicast transmission 1648 instead of multicast. By "echoing" this information to the receiver 1649 set, suppression of feedback can be achieved even when receivers are 1650 unicasting that feedback instead of multicasting it among the group 1651 [13]. 1652 0 1 2 3 1653 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 1654 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1655 |version| type=3| hdr_len | sequence | 1656 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1657 | source_id | 1658 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1659 | instance_id | grtt |backoff| gsize | 1660 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1661 | flavor = 5 | flags | reserved | 1662 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1663 | header extensions (if applicable) | 1664 | ... | 1665 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1666 | repair_adv_payload | 1667 | ... | 1668 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1670 NORM_CMD(REPAIR_ADV) Message Format 1672 The "instance_id", "grtt", "backoff", "gsize", and "flavor" fields serve 1673 the same purpose as in other NORM_CMD messages. The value of the 1674 "hdr_len" field when no extensions are present is 4. 1676 The "flags" field provide information on the NORM_CMD(REPAIR_ADV) 1677 content. There is currently one NORM_CMD(REPAIR_ADV) flag defined: 1679 NORM_REPAIR_ADV_FLAG_LIMIT = 0x01 1681 This flag is set by the sender when it is unable to fit its full current 1682 repair state into a single NormSegmentSize. If this flag is set, 1683 receivers should limit their NACK response to generating NACK content 1684 only up through the maximum ordinal transmission position 1685 (objectId::fecPayloadId) included in the "repair_adv_content". 1687 When congestion control operation is enabled, a header extension may be 1688 applied to the NORM_CMD(REPAIR_ADV) representing the most limiting (in 1689 of congestion control feedback suppression) congestion control response. 1690 This allows the NORM_CMD(REPAIR_ADV) message to suppress receiver 1691 congestion control responses as well as NACK feedback messages. The 1692 field is defined as a header extension so that alternative congestion 1693 control schemes may be used with NORM without revision to this document. 1694 A NORM-CC Feedback Header Extension (EXT_CC) is defined to encapsulate 1695 congestion control feedback within NORM_NACK, NORM_ACK, and 1696 NORM_CMD(REPAIR_ADV) messages. If another congestion control technique 1697 (e.g., Pragmatic General Multicast Congestion Control (PGMCC) [20]) is 1698 used within a NORM implementation, an additional header extension MAY 1699 need to be defined encapsulate any required feedback content. The NORM- 1700 CC Feedback Header Extension format is: 1702 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1703 | ext_type = 3 | ext_len = 3 | cc_sequence | 1704 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1705 | cc_flags | cc_rtt | cc_loss | 1706 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1707 | cc_rate | cc_reserved | 1708 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1710 NORM-CC Feedback Header Extension (EXT_CC) Format 1712 The "cc_sequence" field contains the current greatest "cc_sequence" 1713 value receivers have received in NORM_CMD(CC) messages from the sender. 1714 This information assists the sender in congestion control operation by 1715 providing an indicator of how current ("fresh") the receiver's round- 1716 trip measurement reference time is and whether the receiver has been 1717 successfully receiving recent congestion control probes. For example, 1718 if it is apparent the receiver has not been receiving recent congestion 1719 control probes (and thus possibly other messages from the sender), the 1720 sender may choose to take congestion avoidance measures. For 1721 NORM_CMD(REPAIR_ADV) messages, the sender SHALL set the "cc_sequence" 1722 field value to the value set in the last NORM_CMD(CC) message sent. 1724 The "cc_flags" field contains bits representing the receiver's state 1725 with respect to congestion control operation. The possible values for 1726 the "cc_flags" field are those specified for the NORM_CMD(CC) message 1727 node list item flags. These fields are used by receivers in controlling 1728 (suppressing as necessary) their congestion control feedback. For 1729 NORM_CMD(REPAIR_ADV) messages, the NORM_FLAG_CC_RTT should be set only 1730 when all feedback messages received by the sender have the flag set. 1731 Similarly, the NORM_FLAG_CC_CLR or NORM_FLAG_CC_PLR should be set only 1732 when no feedback has been received from non-CLR or non-PLR receivers. 1733 And the NORM_FLAG_CC_LEAVE should be set only when all feedback messages 1734 the sender has received have this flag set. These heuristics for 1735 setting the flags in NORM_CMD(REPAIR_ADV) ensure the most effective 1736 suppression of receivers providing unicast feedback messages. 1738 The "cc_rtt" field SHALL be set to a default maximum value and the 1739 NORM_FLAG_CC_RTT flag SHALL be cleared when no receiver has yet received 1740 RTT measurement information. When a receiver has received RTT 1741 measurement information, it shall set the "cc_rtt" value accordingly and 1742 the NORM_FLAG_CC_RTT flag in the "cc_flags" field. For 1743 NORM_CMD(REPAIR_ADV) messages, the sender SHALL set the "cc_rtt" field 1744 value to the largest non-CLR/non-PLR RTT it has measured from receivers 1745 for the current feedback round. 1747 The "cc_loss" field represents the receiver's current packet loss 1748 fraction estimate for the indicated source. The loss fraction is a 1749 value from 0.0 to 1.0 corresponding to a range of zero to 100 percent 1750 packet loss. The 16-bit "cc_loss" value is calculated by the following 1751 formula: 1753 "cc_loss" = decimal_loss_fraction * 65535.0 1755 For NORM_CMD(REPAIR_ADV) messages, the sender SHALL set the "cc_loss" 1756 field value to the largest non-CLR/non-PLR loss estimate it has received 1757 from receivers for the current feedback round. 1759 The "cc_rate" field represents the receivers current local congestion 1760 control rate. During "slow start", when the receiver has detected no 1761 loss, this value is set to twice the actual rate it has measured from 1762 the corresponding sender and the NORM_FLAG_CC_START is set in the 1763 "cc_flags' field. Otherwise, the receiver calculates a congestion 1764 control rate based on its loss measurement and RTT measurement 1765 information (even if default) for the "cc_rate" field. For 1766 NORM_CMD(REPAIR_ADV) messages, the sender SHALL set the "cc_loss" field 1767 value to the lowest non-CLR/non-PLR "cc_rate" report it has received 1768 from receivers for the current feedback round. 1770 The "cc_reserved" field is reserved for future NORM protocol use. 1771 Currently, senders SHALL set this field to ZERO, and receivers SHALL 1772 ignore the content of this field. 1774 The "repair_adv_payload" is in exactly the same form as the 1775 "nack_content" of NORM_NACK messages and can be processed by receivers 1776 for suppression purposes in the same manner, with the exception of the 1777 condition when the NORM_REPAIR_ADV_FLAG_LIMIT is set. 1779 4.2.3.6. NORM_CMD(ACK_REQ) Message 1781 The NORM_CMD(ACK_REQ) message is used by the sender to request 1782 acknowledgment from a specified list of receivers. This message is used 1783 in providing a lightweight positive acknowledgment mechanism that is 1784 OPTIONAL for use by the reliable multicast application. A range of 1785 acknowledgment request types is provided for use at the application's 1786 discretion. Provision for application-defined, positively-acknowledged 1787 commands allows the application to automatically take advantage of 1788 transmission and round-trip timing information available to the NORM 1789 protocol. The details of the NORM positive acknowledgment process 1790 including transmission of the NORM_CMD(ACK_REQ) messages and the 1791 receiver response (NORM_ACK) are described in Section 5.5.3. The 1792 format of the NORM_CMD(ACK_REQ) message is: 1794 0 1 2 3 1795 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 1796 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1797 |version| type=3| hdr_len | sequence | 1798 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1799 | source_id | 1800 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1801 | instance_id | grtt |backoff| gsize | 1802 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1803 | flavor = 6 | reserved | ack_type | ack_id | 1804 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1805 | acking_node_list | 1806 | ... | 1807 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1809 NORM_CMD(ACK_REQ) Message Format 1811 The NORM common message header and standard NORM_CMD fields serve their 1812 usual purposes. The value of the "hdr_len" field for NORM_CMD(ACK_REQ) 1813 messages with no header extension present is 4. 1815 The "ack_type" field indicates the type of acknowledgment being 1816 requested and thus implies rules for how the receiver will treat this 1817 request. The following "ack_type" values are defined and are also used 1818 in NORM_ACK messages described later: 1820 +---------------------+--------+----------------------------------+ 1821 | ACK Type | Value | Purpose | 1822 +---------------------+--------+----------------------------------+ 1823 |NORM_ACK_CC | 1 | Used to identify NORM_ACK | 1824 | | | messages sent in response to | 1825 | | | NORM_CMD(CC) messages. | 1826 +---------------------+--------+----------------------------------+ 1827 |NORM_ACK_FLUSH | 2 | Used to identify NORM_ACK | 1828 | | | messages sent in response to | 1829 | | | NORM_CMD(FLUSH) messages. | 1830 +---------------------+--------+----------------------------------+ 1831 |NORM_ACK_RESERVED | 3-15 | Reserved for possible future | 1832 | | | NORM protocol use. | 1833 +---------------------+--------+----------------------------------+ 1834 |NORM_ACK_APPLICATION | 16-255 | Used at application's | 1835 | | | discretion. | 1836 +---------------------+--------+----------------------------------+ 1838 The NORM_ACK_CC value is provided for use only in NORM_ACKs generated in 1839 response to the NORM_CMD(CC) messages used in congestion control 1840 operation. Similarly, the NORM_ACK_FLUSH is provided for use only in 1841 NORM_ACKs generated in response to applicable NORM_CMD(FLUSH) messages. 1842 NORM_CMD(ACK_REQ) messages with "ack_type" of NORM_ACK_CC or 1843 NORM_ACK_FLUSH SHALL NOT be generated by the sender. 1845 The NORM_ACK_RESERVED range of "ack_type" values is provided for 1846 possible future NORM protocol use. 1848 The NORM_ACK_APPLICATION range of "ack_type" values is provided so that 1849 NORM applications may implement application-defined, positively- 1850 acknowledged commands that are able to leverage internal transmission 1851 and round-trip timing information available to the NORM protocol 1852 implementation. 1854 The "ack_id" provides a sequenced identifier for the given 1855 NORM_CMD(ACK_REQ) message. This "ack_id" is returned in NORM_ACK 1856 messages generated by the receivers so that the sender may associate the 1857 response with its corresponding request. 1859 The "reserved" field is reserved for possible future protocol use and 1860 SHALL be set to ZERO by senders and ignored by receivers. 1862 The "acking_node_list" field contains the NormNodeIds of the current 1863 NORM receivers that are desired to provide positive acknowledge 1864 (NORM_ACK) to this request. The packet payload length implies the 1865 length of the "acking_node_list" and its length is limited to the sender 1866 NormSegmentSize. The individual NormNodeId items are listed in network 1867 (Big Endian) byte order. If a receiver's NormNodeId is included in the 1868 "acking_node_list", it SHALL schedule transmission of a NORM_ACK message 1869 as described in Section 5.5.3. 1871 4.2.3.7. NORM_CMD(APPLICATION) Message 1873 This command allows the NORM application to robustly transmit 1874 application-defined commands. The command message preempts any ongoing 1875 data transmission and is repeated up to NORM_ROBUST_FACTOR times at a 1876 rate of once per 2*GRTT. This rate of repetition allows the application 1877 to observe any response (if that is the application's purpose for the 1878 command) before it is repeated. Possible responses may include 1879 initiation of data transmission , other NORM_CMD(APPLICATION) messages, 1880 or even application-defined, positively-acknowledge commands from other 1881 NormSession participants. The transmission of these commands will 1882 preempt data transmission when they are scheduled and may be multiplexed 1883 with ongoing data transmission. This type of robustly transmitted 1884 command allows NORM applications to define a complete set of session 1885 control mechanisms with less state than the transfer of FEC encoded 1886 reliable content requires while taking advantage of NORM transmission 1887 and round-trip timing information. 1889 0 1 2 3 1890 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 1891 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1892 |version| type=3| hdr_len | sequence | 1893 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1894 | source_id | 1895 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1896 | instance_id | grtt |backoff| gsize | 1897 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1898 | flavor = 7 | reserved | 1899 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1900 | Application-Defined Content | 1901 | ... | 1902 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1904 NORM_CMD(APPLICATION) Message Format 1906 The NORM common message header and NORM_CMD fields are interpreted as 1907 previously described. The value of the NORM_CMD(APPLICATION) "hdr_len" 1908 field when no header extensions are present is 4. 1910 The "Application-Defined Content" area contains information in a format 1911 at the discretion of the application. The size of this payload SHALL be 1912 limited to a maximum of the sender's NormSegmentSize setting. 1914 4.3. Receiver Messages 1916 The NORM message types generated by participating receivers consist of 1917 NORM_NACK and NORM_ACK message types. NORM_NACK messages are sent to 1918 request repair of missing data content from sender transmission and 1919 NORM_ACK messages are generated in response to certain sender commands 1920 including NORM_CMD(CC) and NORM_CMD(ACK_REQ). 1922 4.3.1. NORM_NACK Message 1924 The principal purpose of NORM_NACK messages is for receivers to request 1925 repair of sender content via selective, negative acknowledgment upon 1926 detection of incomplete data. NORM_NACK messages will be transmitted 1927 according to the rules of NORM_NACK generation and suppression described 1928 in Section 5.3. NORM_NACK messages also contain additional fields to 1929 provide feedback to the sender(s) for purposes of round-trip timing 1930 collection and congestion control. 1932 The payload of NORM_NACK messages contains one or more repair requests 1933 for different objects or portions of those objects. The NORM_NACK 1934 message format is as follows: 1936 0 1 2 3 1937 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 1938 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1939 |version| type=4| hdr_len | sequence | 1940 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1941 | source_id | 1942 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1943 | server_id | 1944 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1945 | instance_id | reserved | 1946 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1947 | grtt_response_sec | 1948 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1949 | grtt_response_usec | 1950 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1951 | header extensions (if applicable) | 1952 | ... | 1953 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1954 | nack_payload | 1955 | ... | 1956 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1958 NORM_NACK Message Format 1960 The NORM common message header fields serve their usual purposes. The 1961 value of the "hdr_len" field for NORM_NACK messages without header 1962 extensions present is 6. 1964 The "server_id" field identifies the NORM sender to which the NORM_NACK 1965 message is destined. 1967 The "instance_id" field contains the current session identifier given by 1968 the sender identified by the "server_id" field in its sender messages. 1969 The sender SHOULD ignore feedback messages which contain an invalid 1970 "instance_id" value. 1972 The "grtt_response" fields contain an adjusted version of the timestamp 1973 from the most recently received NORM_CMD(CC) message for the indicated 1974 NORM sender. The format of the "grtt_response" is the same as the 1975 "send_time" field of the NORM_CMD(CC). The "grtt_response" value is 1976 _relative_ to the "send_time" the source provided with a corresponding 1977 NORM_CMD(CC) command. The receiver adjusts the source's NORM_CMD(CC) 1978 "send_time" timestamp by adding the time differential from when the 1979 receiver received the NORM_CMD(CC) to when the NORM_NACK is transmitted 1980 to calculate the value in the "grtt_response" field. This is the 1981 "receive_to_response_differential" value used in the following formula: 1983 "grtt_response" = NORM_CMD(CC) "send_time" + receive_to_response_differential 1985 The receiver SHALL set the "grtt_response" to a ZERO value, to indicate 1986 that it has not yet received a NORM_CMD(CC) message from the indicated 1987 sender and that the sender should ignore the "grtt_response" in this 1988 message. 1990 For NORM-CC operation, the NORM-CC Feedback Header Extension, as 1991 described in the NORM_CMD(REPAIR_ADV} message description, is added to 1992 NORM_NACK messages to provide feedback on the receivers current state 1993 with respect to congestion control operation. Note that alternative 1994 header extensions for congestion control feedback may be defined for 1995 alternative congestion control schemes for NORM use in the future. 1997 The "reserved" field is for potential future NORM use and SHALL be set 1998 to ZERO for this version of the protocol. 2000 The "nack_content" of the NORM_NACK message specifies the repair needs 2001 of the receiver with respect to the NORM sender indicated by the 2002 "server_id" field. The receiver constructs repair requests based on the 2003 NORM_DATA and/or NORM_INFO segments it requires from the sender in order 2004 to complete reliable reception up to the sender's transmission position 2005 at the moment the receiver initiates the NACK Procedure as described in 2006 Section 5.3. A single NORM Repair Request consists of a list of items, 2007 ranges, and/or FEC coding block erasure counts for needed NORM_DATA 2008 and/or NORM_INFO content. Multiple repair requests may be concatenated 2009 within the "nack_payload" field of a NORM_NACK message. Note that a 2010 single NORM Repair Request can possibly include multiple "items", 2011 "ranges", or "erasure_counts". In turn, the "nack_payload" field may 2012 contain multiple repair requests. A single NORM Repair Request has the 2013 following format: 2015 0 1 2 3 2016 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2017 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2018 | form | flags | length | 2019 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2020 | repair_request_items | 2021 | ... | 2022 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2024 NORM Repair Request Format 2026 The "form" field indicates the type of repair request items given in the 2027 "repair_request_items" list. Possible values for the "form" field 2028 include: 2030 Form Value 2031 NORM_NACK_ITEMS 1 2032 NORM_NACK_RANGES 2 2033 NORM_NACK_ERASURES 3 2035 A "form" value of NORM_NACK_ITEMS indicates each repair request item in 2036 the "repair_request_items" list is to be treated as an individual 2037 request. A value of NORM_NACK_RANGES indicates that the 2038 "repair_request_items" list consists of pairs of repair request items 2039 that correspond to inclusive ranges of repair needs. And the 2040 NORM_NACK_ERASURES "form" indicates that the repair request items are to 2041 be treated individually and that the "encoding_symbol_id" portion of the 2042 field of the repair request item (see below) is to be interpreted as an 2043 "erasure count" for the FEC coding block identified by the repair 2044 request item's "source_block_number". 2046 The "flags" field is currently used to indicate the level of data 2047 content for which the repair request items apply (i.e., an individual 2048 segment, entire FEC coding block, or entire transport object). Possible 2049 flag values include: 2051 +------------------+-------+------------------------------------------+ 2052 | Flag | Value | Purpose | 2053 +------------------+-------+------------------------------------------+ 2054 |NORM_NACK_SEGMENT | 0x01 | Indicates the listed segment(s) or range | 2055 | | | of segments are required as repair. | 2056 +------------------+-------+------------------------------------------+ 2057 |NORM_NACK_BLOCK | 0x02 | Indicates the listed block(s) or range | 2058 | | | of blocks in entirety are required as | 2059 | | | repair. | 2060 +------------------+-------+------------------------------------------+ 2061 |NORM_NACK_INFO | 0x04 | Indicates that NORM_INFO is required as | 2062 | | | repair for the listed object(s). | 2063 +------------------+-------+------------------------------------------+ 2064 |NORM_NACK_OBJECT | 0x08 | Indicates the listed object(s) or range | 2065 | | | of objects in entirety are required as | 2066 | | | repair. | 2067 +------------------+-------+------------------------------------------+ 2069 When the NORM_NACK_SEGMENT flag is set, the "object_transport_id" and 2070 "fec_payload_id" fields are used to determine which sets or ranges of 2071 individual NORM_DATA segments are needed to repair content at the 2072 receiver. When the NORM_NACK_BLOCK flag is set, this indicates the 2073 receiver is completely missing the indicated coding block(s) and 2074 requires transmissions sufficient to repair the indicated block(s) in 2075 their entirety. When the NORM_NACK_INFO flag is set, this indicates the 2076 receiver is missing the NORM_INFO segment for the indicated 2077 "object_transport_id". Note the NORM_NACK_INFO may be set in 2078 combination with the NORM_NACK_BLOCK or NORM_NACK_SEGMENT flags, or may 2079 be set alone. When the NORM_NACK_OBJECT flag is set, this indicates the 2080 receiver is missing the entire NormTransportObject referenced by the 2081 "object_transport_id". This also implicitly requests any available 2082 NORM_INFO for the NormObject, if applicable. The "fec_payload_id" field 2083 is ignored when the flag NORM_NACK_OBJECT is set. 2085 The "length" field value is the length in bytes of the 2086 "repair_request_items" field. 2088 The "repair_request_items" field consists of a list of individual or 2089 range pairs of transport data unit identifiers in the following format. 2091 0 1 2 3 2092 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2093 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2094 | fec_id | reserved | object_transport_id | 2095 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2096 | fec_payload_id | 2097 | ... | 2098 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2100 NORM Repair Request Item Format 2102 The "fec_id" indicates the FEC type and can be used to determine the 2103 format of the "fec_payload_id" field. The "reserved" field is kept for 2104 possible future use and SHALL be set to a ZERO value and ignored by NORM 2105 nodes processing NACK content. 2107 The "object_transport_id" corresponds to the NormObject for which repair 2108 is being requested and the "fec_payload_id" identifies the specific FEC 2109 coding block and/or segment being requested. When the NORM_NACK_OBJECT 2110 flag is set, the value of the "fec_payload_id" field is ignored. When 2111 the NORM_NACK_BLOCK flag is set, only the FEC code block identifier 2112 portion of the "fec_payload_id" is to be interpreted. 2114 The format of the "fec_payload_id" field depends upon the "fec_id" field 2115 value. 2117 When the receiver's repair needs dictate that different forms (mixed 2118 ranges and/or individual items) or types (mixed specific segments and/or 2119 blocks or objects in entirety) are required to complete reliable 2120 transmission, multiple NORM Repair Requests with different "form" and or 2121 "flags" values can be concatenated within a single NORM_NACK message. 2122 Additionally, NORM receivers SHALL construct NORM_NACK messages with 2123 their repair requests in ordinal order with respect to 2124 "object_transport_id" and "fec_payload_id" values. The "nack_payload" 2125 size SHALL NOT exceed the NormSegmentSize for the sender to which the 2126 NORM_NACK is destined. 2128 NORM_NACK Content Examples: 2130 In these examples, a small block, systematic FEC code ("fec_id" = 129) 2131 is assumed with a user data block length of 32 segments. In Example 1, 2132 a list of individual NORM_NACK_ITEMS repair requests is given. In 2133 Example 2, a list of NORM_NACK_RANGES requests _and_ a single 2134 NORM_NACK_ITEMS request are concatenated to illustrate the possible 2135 content of a NORM_NACK message. Note that FEC coding block erasure 2136 counts could also be provided in each case. However, the erasure counts 2137 are not really necessary since the sender can easily determine the 2138 erasure count while processing the NACK content. However, the erasure 2139 count option may be useful for operation with other FEC codes or for 2140 intermediate system purposes. 2142 Example 1: NORM_NACK "nack_payload" for: Object 12, Coding Block 3, Segments 2,5,8 2143 0 1 2 3 2144 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2145 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2146 | form = 1 | flags = 0x01 | length = 36 | 2147 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2148 | fec_id = 129 | reserved | object_transport_id = 12 | 2149 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2150 | source_block_number = 3 | 2151 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2152 | source_block_length = 32 | encoding_symbol_id = 2 | 2153 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2154 | fec_id = 129 | reserved | object_transport_id = 12 | 2155 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2156 | source_block_number = 3 | 2157 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2158 | source_block_length = 32 | encoding_symbol_id = 5 | 2159 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2160 | fec_id = 129 | reserved | object_transport_id = 12 | 2161 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2162 | source_block_number = 3 | 2163 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2164 | source_block_length = 32 | encoding_symbol_id = 8 | 2165 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2167 Example 2: NORM_NACK "nack_payload" for: Object 18 Coding Block 6, 2168 Segments 5, 6, 7, 8, 9, 10; and Object 19 NORM_INFO and Coding Block 1, 2169 segment 3 2170 0 1 2 3 2171 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2172 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2173 | form = 2 | flags = 0x01 | length = 24 | 2174 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2175 | fec_id = 129 | reserved | object_transport_id = 18 | 2176 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2177 | source_block_number = 6 | 2178 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2179 | source_block_length = 32 | encoding_symbol_id = 5 | 2180 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2181 | fec_id = 129 | reserved | object_transport_id = 18 | 2182 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2183 | source_block_number = 6 | 2184 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2185 | source_block_length = 32 | encoding_symbol_id = 10 | 2186 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2187 | form = 1 | flags = 0x05 | length = 12 | 2188 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2189 | fec_id = 129 | reserved | object_transport_id = 19 | 2190 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2191 | source_block_number = 1 | 2192 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2193 | source_block_length = 32 | encoding_symbol_id = 3 | 2194 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2195 4.3.2. NORM_ACK Message 2197 The NORM_ACK message is intended to be used primarily as part of NORM 2198 congestion control operation and round-trip timing measurement. As 2199 mentioned in the NORM_CMD(ACK_REQ) message description, the 2200 acknowledgment type NORM_ACK_CC is provided for this purpose. The 2201 generation of NORM_ACK(CC) messages for round-trip timing estimation and 2202 congestion-control operation is described in Sections 5.5.1 and 5.5.2, 2203 respectively. However, some multicast applications may benefit from 2204 some limited form of positive acknowledgment for certain functions. A 2205 simple, scalable positive acknowledgment scheme is defined in Section 2206 5.5.3 that can be leveraged by protocol implementations when 2207 appropriate. The NORM_CMD(FLUSH) may be used for OPTIONAL collection of 2208 positive acknowledgment of reliable reception to a certain "watermark" 2209 transmission point from specific receivers using this mechanism. The 2210 NORM_ACK type NORM_ACK_FLUSH is provided for this purpose and the format 2211 of the "nack_payload" for this acknowledgment type is given below. 2212 Beyond that, a range of application-defined "ack_type" values is 2213 provided for use at the NORM application's discretion. Implementations 2214 making use of application-defined positive acknowledgments may also make 2215 use the "nack_payload" as needed, observing the constraint that the 2216 "nack_payload" field size be limited to a maximum of the NormSegmentSize 2217 for the sender to which the NORM_ACK is destined. 2219 0 1 2 3 2220 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2221 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2222 |version| type=5| hdr_len | sequence | 2223 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2224 | source_id | 2225 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2226 | server_id | 2227 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2228 | instance_id | ack_type | ack_id | 2229 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2230 | grtt_response_sec | 2231 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2232 | grtt_response_usec | 2233 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2234 | header extensions (if applicable) | 2235 | ... | 2236 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2237 | ack_payload (if applicable) | 2238 | ... | 2239 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2241 NORM_ACK Message Format 2243 The NORM common message header fields serve their usual purposes. 2245 The "server_id", "instance_id", and "grtt_response" fields serve the 2246 same purpose as the corresponding fields in NORM_NACK messages. And 2247 header extensions may be applied to support congestion control feedback 2248 other functions in the same manner. 2250 The "ack_type" field indicates the nature of the NORM_ACK message. This 2251 directly corresponds to the "ack_type" field of the NORM_CMD(ACK_REQ) 2252 message to which this acknowledgment applies. 2254 The "ack_id" field serves as a sequence number so that the sender can 2255 verify that a NORM_ACK message received actually applies to a current 2256 acknowledgment request. The "ack_id" field is not used in the case of 2257 the NORM_ACK_CC and NORM_ACK_FLUSH acknowledgment types. 2259 The "ack_payload" format is a function of the "ack_type". The 2260 NORM_ACK_CC message has no attached content. Only the NORM_ACK header 2261 applies. In the case of NORM_ACK_FLUSH, a specific "ack_payload" format 2262 is defined: 2264 0 1 2 3 2265 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2266 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2267 | fec_id | reserved | object_transport_id | 2268 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2269 | fec_payload_id | 2270 | ... | 2271 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2273 NORM_ACK_FLUSH "ack_payload" Format 2275 The "object_transport_id" and "fec_payload_id" are used by the receiver 2276 to acknowledge applicable NORM_CMD(FLUSH) messages transmitted by the 2277 sender identified by the "server_id" field. 2279 The "ack_payload" of NORM_ACK messages for application-defined 2280 "ack_type" values is specific to the application but is limited in size 2281 to a maximum the NormSegmentSize of the sender referenced by the 2282 "server_id". 2284 4.4. General Purpose Messages 2286 Some additional message formats are defined for general purpose in NORM 2287 multicast sessions whether the participant is acting as a sender and/or 2288 receiver within the group. 2290 4.4.1. NORM_REPORT Message 2292 This is an optional message generated by NORM participants. This 2293 message could be used for periodic performance reports from receivers in 2294 experimental NORM implementations. The format of this message is 2295 currently undefined. Experimental NORM implementations may define 2296 NORM_REPORT formats as needed for test purposes. These report messages 2297 SHOULD be disabled for interoperability testing between different NORM 2298 implementations. 2300 5. Detailed Protocol Operation 2302 This section describes the detailed interactions of senders and 2303 receivers participating in a NORM session. A simple synopsis of 2304 protocol operation is given here: 2306 1) The sender periodically transmits NORM_CMD(CC) messages as 2307 needed to initialize and collect roundtrip timing and 2308 congestion control feedback from the receiver set. 2310 2) The sender transmits an ordinal set of NormObjects segmented 2311 in the form of NORM_DATA messages labeled with 2312 NormTransportIds and logically identified with FEC encoding 2313 block numbers and symbol identifiers. NORM_INFO messages 2314 may optionally precede the transmission of data content for 2315 NORM transport objects. 2317 3) As receivers detect missing content from the sender, they 2318 initiate repair requests with NORM_NACK messages. Note the 2319 receivers track the sender's most recent 2320 objectId::fecPayloadId transmit position and NACK _only_ for 2321 content ordinally prior to that transmit position. The 2322 receivers schedule random backoff timeouts before generating 2323 NORM_NACK messages and wait an appropriate amount of time 2324 before repeating the NORM_NACK if their repair request is 2325 not satisfied. 2327 4) The sender aggregates repair requests from the receivers and 2328 logically "rewinds" its transmit position to send 2329 appropriate repair messages. The sender sends repairs for 2330 the earliest ordinal transmit position first and maintains 2331 this ordinal repair transmission sequence. Previously 2332 untransmitted FEC parity content for the applicable FEC 2333 coding block is used for repair transmissions to the 2334 greatest extent possible. If the sender exhausts its 2335 available FEC parity content on multiple repair cycles for 2336 the same coding block, it resorts to an explicit repair 2337 strategy (possibly using parity content) to complete 2338 repairs. (The use of explicit repair is expected to be an 2339 exception in general protocol operation, but the possibility 2340 does exist for extreme conditions). The sender immediately 2341 assumes transmission of new content once it has sent pending 2342 repairs. 2344 5) The sender transmits NORM_CMD(FLUSH) messages when it 2345 reaches the end of enqueued transmit content and pending 2346 repairs. Receivers respond to the NORM_CMD(FLUSH) messages 2347 with NORM_NACK transmissions (following the same suppression 2348 backoff timeout strategy as for data) if they require 2349 further repair. 2351 6) The sender transmissions are subject to rate control limits 2352 determined by congestion control mechanisms. In the 2353 baseline NORM-CC operation, each sender in a NormSession 2354 maintains its own independent congestion control state. 2355 Receivers provide congestion control feedback in NORM_NACK 2356 and NORM_ACK messages. NORM_ACK feedback for congestion 2357 control purposes is governed using a suppression mechanism 2358 similar to that for NORM_NACK messages. 2360 While this overall concept is relatively simple, there are details to 2361 each of these aspects that need to be addressed for successful, 2362 efficient, robust, and scalable NORM protocol operation. 2364 5.1. Sender Initialization and Transmission 2366 Upon startup, the NORM sender immediately begins sending NORM_CMD(CC) 2367 messages to collect round trip timing and other information from the 2368 potential group. If NORM-CC congestion control operation is enabled, 2369 the NORM-CC Rate header extension MUST be included in these messages. 2370 Congestion control operation SHALL be observed at all times when 2371 operating in the general Internet. Even if congestion control operation 2372 is disabled at the sender, it may be desirable to use the NORM_CMD(CC) 2373 messaging to collect feedback from the group using the baseline NORM-CC 2374 feedback mechanisms. This proactive feedback collection can be used to 2375 establish a GRTT estimate prior to data transmission and potential NACK 2376 operation. 2378 In some cases, applications may wish for the sender to also proceed with 2379 data transmission immediately. In other cases, the sender may wish to 2380 defer data transmission until it has received some feedback or request 2381 from the receiver set indicating that receivers are indeed present. 2382 Note, in some applications (e.g., web push), this indication may come 2383 out-of-band with respect to the multicast session via other means. As 2384 noted, the periodic transmission of NORM_CMD(CC) messages may precede 2385 actual data transmission in order to have an initial GRTT estimate. 2387 With inclusion of the OPTIONAL NORM FEC Object Transmission Information 2388 Header Extension, the NORM protocol sender message headers can contain 2389 all information necessary to prepare receivers for subsequent reliable 2390 reception. This includes FEC coding parameters, the sender 2391 NormSegmentSize, and other information. If this header extension is not 2392 used, it is presumed that receivers have received the FEC Object 2393 Transmission Information via other means. Additionally, applications 2394 may leverage the use of NORM_INFO messages associated with the session 2395 data objects in the session to provide application-specific context 2396 information for the session and data being transmitted. These 2397 mechanisms allow for operation with minimal pre-coordination among the 2398 senders and receivers. 2400 The NORM sender begins segmenting application-enqueued data into 2401 NORM_DATA segments and transmitting it to the group. The segmentation 2402 algorithm is described in Section 5.1.1. The rate of transmission is 2403 via congestion control mechanisms or is a fixed rate if desired for 2404 closed network operations. The receivers participating in the multicast 2405 group provide feedback to the sender as needed. When the sender reaches 2406 the end of data it has enqueued for transmission or any pending repairs, 2407 it transmits a series of NORM_CMD(FLUSH) messages at a rate of one per 2408 2*GRTT. Receivers may respond to these NORM_CMD(FLUSH) messages with 2409 additional repair requests. A protocol parameter "NORM_ROBUST_FACTOR" 2410 determines the number of flush messages sent. If receivers request 2411 repair, the repair is provided and flushing occurs again at the end of 2412 repair transmission. The sender may attach an OPTIONAL 2413 "acking_node_list" to NORM_CMD(FLUSH) containing the NormNodeIds for 2414 receivers from which it expects explicit positive acknowledgment of 2415 reception. The NORM_CMD(FLUSH) message may be also used for this 2416 optional function any time prior to the end of data enqueued for 2417 transmission with the NORM_CMD(FLUSH) messages multiplexed with ongoing 2418 data transmissions. The OPTIONAL NORM positive acknowledgment procedure 2419 is described in Section 5.5.3. 2421 5.1.1. Object Segmentation Algorithm 2423 NORM senders and receivers must use a common algorithm for logically 2424 segmenting transport data into FEC encoding blocks and symbols so that 2425 appropriate NACKs can be constructed to request repair of missing data. 2426 NORM FEC coding blocks are comprised of multi-byte symbols which are 2427 transmitted in the payload of NORM_DATA messages. Each NORM_DATA 2428 message contains one source or encoding symbol and the NormSegmentSize 2429 sender parameter defines the maximum symbol size in bytes. The FEC 2430 encoding type and associated parameters govern the source block size 2431 (number of source symbols per coding block). NORM senders and receivers 2432 use these FEC parameters, along with the NormSegmentSize and transport 2433 object size to compute the source block structure for transport objects. 2434 These parameters are provided in the FEC Transmission Information for 2435 each object. The algorithm given below is used to compute a source 2436 block structure such that all source blocks are as close to being equal 2437 length as possible. This helps avoid the performance disadvantages of 2438 "short" FEC blocks. Note this algorithm applies only to the statically- 2439 sized NORM_OBJECT_DATA and NORM_OBJECT_FILE transport object types where 2440 the object size is fixed and predetermined. For NORM_OBJECT_STREAM 2441 objects, the object is segmented according to the maximum source block 2442 length given in the FEC Transmission Information, unless the FEC 2443 Payload ID indicates an alternative size for a given block. 2445 The NORM block segmentation algorithm is defined as follows. For a 2446 transport object of a given length (L_obj) in bytes , a first number of 2447 FEC source blocks (N_large) is delineated of a larger block size 2448 (B_large), and a second number of source blocks (N_small) is delineated 2449 of a smaller block size (B_small). Given the maximum FEC source block 2450 size (B_max) and the sender's NormSegmentSize, the block segmentation 2451 for a given NORM transport object is determined as follows: 2453 Inputs: 2455 B_max = Maximum source block length (i.e., maximum number of source 2456 symbols per source block) 2458 L_sym = Encoding symbol length in bytes (i.e., NormSegmentSize) 2460 L_obj = Object length in bytes 2462 Outputs: 2464 N_total = The total number of source blocks into which the transport 2465 object is partitioned. 2467 N_large = Number of larger source blocks (first set of blocks) 2469 B_large = Size (in encoding symbols) of the larger source blocks 2471 N_small = Number of smaller source blocks (second set of blocks) 2473 B_small = Size (in encoding symbols) of the smaller source blocks 2475 L_final = Length (in bytes) of the last source symbol of the last 2476 source block (All other symbols are of length L_sym). 2478 Algorithm: 2480 1) The total number of source symbols in the transport object is computed as: 2481 S_total = L_obj/L_sym [rounded up to the nearest integer] 2483 2) The transport object is partitioned into N_total source blocks, where: 2484 N_total = S_total/B_max [rounded up to the nearest integer] 2486 3) The average length of a source block is computed as: 2487 B_ave = S_total/N_total (this may be non-integer) 2489 4) The size of the first set of larger blocks is computed as: 2490 B_large = B_ave [rounded up to the nearest integer] 2491 (Note it will always be the case that B_large <= B_max) 2493 5) The size of the second set of smaller blocks is computed as: 2494 B_small = B_ave [rounded down to the nearest integer] 2495 (Note if B_ave is an integer B_small = B_large; otherwise B_small = B_large 2496 - 1) 2498 6) The fractional part of B_ave is computed as: 2499 B_fraction = B_ave - B_small 2501 7) The number of larger source blocks is computed as: 2502 N_large = B_fraction * N_total 2503 (Note N_large is an integer in the range 0 through N_total - 1) 2505 8) The number of smaller source blocks is computed as: 2507 N_small = N_total - N_large 2509 9) Each of the first N_large source blocks consists of B_large source symbols. 2510 Each of the remaining N_small source blocks consists of B_small source 2511 symbols. All symbols are L_sym bytes in length except for the final source 2512 symbol of the final source block which is of length (in bytes): 2513 L_final = L_obj - (N_large*B_large + N_small*B_small - 1) * L_sym 2515 5.2. Receiver Initialization and Reception 2517 The NORM protocol is designed such that receivers may join and leave the 2518 group at will. However, some applications may be constrained such that 2519 receivers need to be members of the group prior to start of data 2520 transmission. NORM applications may use different policies to constrain 2521 the impact of new receivers joining the group in the middle of a 2522 session. For example, a useful implementation policy is for new 2523 receivers joining the group to limit or avoid repair requests for 2524 transport objects already in progress. The NORM sender implementation 2525 may wish to impose additional constraints to limit the ability of 2526 receivers to disrupt reliable multicast performance by joining, leaving, 2527 and rejoining the group often. Different receiver "join policies" may 2528 be appropriate for different applications and/or scenarios. For general 2529 purpose operation, default policy where receivers are allowed to request 2530 repair only for coding blocks with a NormTransportId and FEC coding 2531 block number greater than or equal to the first non-repair NORM_DATA or 2532 NORM_INFO message received upon joining the group is RECOMMENDED. For 2533 objects of type NORM_OBJECT_STREAM it is RECOMMENDED that the join 2534 policy constrain receivers to start reliable reception at the current 2535 FEC coding block for which non-repair content is received. 2537 5.3. Receiver NACK Procedure 2539 When the receiver detects it is missing data from a sender's NORM 2540 transmissions, it initiates its NACKing procedure. The NACKing 2541 procedure SHALL be initiated _only_ at FEC coding block boundaries, 2542 NormObject boundaries, and upon receipt of a NORM_CMD(FLUSH) message. 2544 The NACKing procedure begins with a random backoff timeout. The 2545 duration of the backoff timeout is chosen using the "RandomBackoff" 2546 algorithm described in the NORM Building Block document [4] using 2547 (Ksender*GRTTsender) for the "maxTime" parameter and the sender 2548 advertised group size (GSIZEsender) as the "groupSize" parameter. NORM 2549 senders provide values for GRTTsender, Ksender and GSIZEsender via the 2550 "grtt", "backoff", and "gsize" fields of transmitted messages. The 2551 GRTTsender value is determined by the sender based on feedback it has 2552 received from the group while the Ksender and GSIZEsender values may 2553 determined by application requirements and expectations or ancillary 2554 information. The backoff factor "Ksender" MUST be greater than one to 2555 provide for effective feedback suppression. A value of K = 4 is 2556 RECOMMENDED for the Any Source Multicast (ASM) model while a value of K 2557 = 6 is RECOMMENDED for Single Source Multicast (SSM) operation. 2559 Thus: 2561 T_backoff = RandomBackoff(Ksender*GRTTsender, GSIZEsender) 2563 To avoid the possibility of NACK implosion in the case of sender or 2564 network failure during SSM operation, the receiver SHALL automatically 2565 suppress its NACK and immediately enter the "holdoff" period described 2566 below when T_backoff is greater than (Ksender-1)*GRTTsender. Otherwise, 2567 the backoff period is entered and the receiver MUST accumulate external 2568 pending repair state from NORM_NACK messages and NORM_CMD(REPAIR_ADV) 2569 messages received. At the end of the backoff time, the receiver SHALL 2570 generate a NORM_NACK message only if the following conditions are met: 2572 1) The sender's current transmit position (in terms of 2573 objectId::fecPayloadId) exceeds the earliest repair position 2574 of the receiver. 2576 2) The repair state accumulated from NORM_NACK and 2577 NORM_CMD(REPAIR_ADV) messages do not equal or supersede the 2578 receiver's repair needs up to the sender transmission 2579 position at the time the NACK procedure (backoff timeout) 2580 was initiated. 2582 If these conditions are met, the receiver immediately generates a 2583 NORM_NACK message when the backoff timeout expires. Otherwise, the 2584 receiver's NACK is considered to be "suppressed" and the message is not 2585 sent. At this time, the receiver begins a "holdoff" period during which 2586 it constrains itself to not reinitiate the NACKing process. The purpose 2587 of this timeout is to allow the sender worst-case time to respond to the 2588 repair needs before the receiver requests repair again. The value of 2589 this "holdoff" timeout (T_rcvrHoldoff) as described in [4] is: 2591 T_rcvrHoldoff =(Ksender+2)*GRTTsender 2593 The NORM_NACK message contains repair request content beginning with 2594 lowest ordinal repair position of the receiver up through the coding 2595 block prior to the most recently heard ordinal transmission position for 2596 the sender. If the size of the NORM_NACK content exceeds the sender's 2597 NormSegmentSize, the NACK content is truncated so that the receiver only 2598 generates a single NORM_NACK message per NACK cycle for a given sender. 2599 In summary, a single NACK message is generated containing the receiver's 2600 lowest ordinal repair needs. 2602 For each partially-received FEC coding block requiring repair, the 2603 receiver SHALL, on its _first_ repair attempt for the block, request the 2604 parity portion of the FEC coding block beginning with the lowest ordinal 2605 _parity_ "encoding_symbol_id" (i.e. "encoding_symbol_id" = 2606 "source_block_len") and request the number of FEC symbols corresponding 2607 to its data segment erasure count for the block. On _subsequent_ repair 2608 cycles for the same coding block, the receiver SHALL request only those 2609 repair symbols from the first set it has not yet received up to the 2610 remaining erasure count for that applicable coding block. Note that the 2611 sender may have provided other different, additional parity segments for 2612 other receivers that could also be used to satisfy the local receiver's 2613 needs. In the case where the erasure count for a partially-received FEC 2614 coding block exceeds the maximum number of parity symbols available from 2615 the sender for the block (as indicated by the NORM_DATA "fec_num_parity" 2616 field), the receiver SHALL request all available parity segments plus 2617 the ordinally highest missing data segments required to satisfy its 2618 total erasure needs for the block. The goal of this strategy is for the 2619 overall receiver set to request a lowest common denominator set of 2620 repair symbols for a given FEC coding block. This allows the sender to 2621 construct the most efficient repair transmission segment set and enables 2622 effective NACK suppression among the receivers even with uncorrelated 2623 packet loss. This approach also requires no synchronization among the 2624 receiver set in their repair requests for the sender. 2626 For FEC coding blocks or NormObjects missed in their entirety, the NORM 2627 receiver constructs repair requests with NORM_NACK_BLOCK or 2628 NORM_NACK_OBJECT flags set as appropriate. The request for 2629 retransmission of NORM_INFO is accomplished by setting the 2630 NORM_NACK_INFO flag in a corresponding repair request. 2632 5.4. Sender NACK Processing and Response 2634 The principle goal of the sender is to make forward progress in the 2635 transmission of data its application has enqueued. However, the sender 2636 must occasionally "rewind" its logical transmission point to satisfy the 2637 repair needs of receivers who have NACKed. Aggregation of multiple 2638 NACKs is used to determine an optimal repair strategy when a NACK event 2639 occurs. Since receivers initiate the NACK process on coding block or 2640 object boundaries, there is some loose degree of synchronization of the 2641 repair process even when receivers experience uncorrelated data loss. 2643 5.4.1. Sender Repair State Aggregation 2645 When a sender is in its normal state of transmitting new data and 2646 receives a NACK, it begins a procedure to accumulate NACK repair state 2647 from NORM_NACK messages before beginning repair transmissions. Note 2648 that this period of aggregating repair state does _not_ interfere with 2649 its ongoing transmission of new data. 2651 As described in [4], the period of time during which the sender 2652 aggregates NORM_NACK messages is equal to: 2654 T_sndrAggregate = (Ksender+1)*GRTT 2656 where "Ksender" is the same backoff scaling value used by the receivers, 2657 and "GRTT" is the sender's current estimate of the group's greatest 2658 round-trip time. 2660 When this period ends, the sender "rewinds" by incorporating the 2661 accumulated repair state into its pending transmission state and begins 2662 transmitting repair messages. After pending repair transmissions are 2663 completed, the sender continues with new transmissions of any enqueued 2664 data. Also, at this point in time, the sender begins a "holdoff" 2665 timeout during which time the sender constrains itself from initiating a 2666 repair aggregation cycle, even if NORM_NACK messages arrive. As 2667 described in [4], the value of this sender "holdoff" period is: 2669 T_sndrHoldoff = (1*GRTT) 2671 If additional NORM_NACK messages are received during this sender 2672 "holdoff" period, the sender will immediately incorporate these "late 2673 messages" into its pending transmission state ONLY if the NACK content 2674 is ordinally greater than the sender's current transmission position. 2675 This "holdoff" time allows worst case time for the sender to propagate 2676 its current transmission sequence position to the group, thus avoiding 2677 redundant repair transmissions. After the holdoff timeout expires, a 2678 new NACK accumulation period can be begun (upon arrival of a NACK) in 2679 concert with the pending repair and new data transmission. Recall that 2680 receivers are not to initiate the NACK repair process until the sender's 2681 logical transmission position exceeds the lowest ordinal position of 2682 their repair needs. With the new NACK aggregation period, the sender 2683 repeats the same process of incorporating accumulated repair state into 2684 its transmission plan and subsequently "rewinding" to transmit the 2685 lowest ordinal repair data when the aggregation period expires. Again, 2686 this is conducted in concert with ongoing new data and/or pending repair 2687 transmissions. 2689 5.4.2. Sender FEC Repair Transmission Strategy 2691 The NORM sender should leverage transmission of FEC parity content for 2692 repair to the greatest extent possible. Recall that the receivers use a 2693 strategy to request a lowest common denominator of explicit repair 2694 (including parity content) in the formation of their NORM_NACK messages. 2695 Before falling back to explicitly satisfying different receivers' repair 2696 needs, the sender can make use of the general erasure-filling capability 2697 of FEC-generated parity segments. The sender can determine the maximum 2698 erasure filling needs for individual FEC coding blocks from the 2699 NORM_NACK messages received during the repair aggregation period. Then, 2700 if the sender has a sufficient number (less than or equal to the maximum 2701 erasure count) of previously unsent parity segments available for the 2702 applicable coding blocks, the sender can transmit these in lieu of the 2703 specific packets the receiver set has requested. Only after exhausting 2704 its supply of "fresh" (unsent) parity segments for a given coding block 2705 should the sender resort to explicit transmission of the receiver set's 2706 repair needs. In general, if a sufficiently powerful FEC code is used, 2707 the need for explicit repair will be an exception, and the fulfillment 2708 of reliable multicast can be accomplished quite efficiently. However, 2709 the ability to resort to explicit repair allows the protocol to be 2710 reliable under even very extreme circumstances. 2712 NORM_DATA messages sent as repair transmissions SHALL be flagged with 2713 the NORM_FLAG_REPAIR flag. This allows receivers to obey any policies 2714 that limit new receivers from joining the reliable transmission when 2715 only repair transmissions have been received. Additionally, the sender 2716 SHOULD additionally flag NORM_DATA transmissions sent as explicit repair 2717 with the NORM_FLAG_EXPLICIT flag. 2719 Although NORM end system receivers do not make use of the 2720 NORM_FLAG_EXPLICIT flag, this message transmission status could be 2721 leveraged by intermediate systems wishing to "assist" NORM protocol 2722 performance. If such systems are properly positioned with respect to 2723 reciprocal reverse-path multicast routing, they need to sub-cast only a 2724 sufficient count of non-explicit parity repairs to satisfy a multicast 2725 routing sub-tree's erasure filling needs for a given FEC coding block. 2726 When the sender has resorted to explicit repair, then the intermediate 2727 systems should sub-cast all of the explicit repair packets to those 2728 portions of the routing tree still requiring repair for a given coding 2729 block. Note the intermediate systems will be required to conduct repair 2730 state accumulation for sub-routes in a manner similar to the sender's 2731 repair state accumulation in order to have sufficient information to 2732 perform the sub-casting. Additionally, the intermediate systems could 2733 perform additional NORM_NACK suppression/aggregation as it conducts this 2734 repair state accumulation for NORM repair cycles. The detail of this 2735 type of operation are beyond the scope of this document, but this 2736 information is provided for possible future consideration. 2738 5.4.3. Sender NORM_CMD(SQUELCH) Generation 2740 If the sender receives a NORM_NACK message for repair of data it is no 2741 longer supporting, the sender generates a NORM_CMD(SQUELCH) message to 2742 advertise its repair window and squelch any receivers from additional 2743 NACKing of invalid data. The transmission rate of NORM_CMD(SQUELCH) 2744 messages is limited to once per 2*GRTT. The "invalid_object_list" (if 2745 applicable) of the NORM_CMD(SQUELCH) message SHALL begin with the lowest 2746 "object_transport_id" from the invalid NORM_NACK messages received since 2747 the last NORM_CMD(SQUELCH) transmission. Lower ordinal invalid 2748 "object_transport_ids" should be included only while the 2749 NORM_CMD(SQUELCH) payload is less than the sender's NormSegmentSize 2750 parameter. 2752 5.4.4. Sender NORM_CMD(REPAIR_ADV) Generation 2754 When a NORM sender receives NORM_NACK messages from receivers via 2755 unicast transmission, it uses NORM_CMD(REPAIR_ADV) messages to advertise 2756 its accumulated repair state to the receiver set since the receiver set 2757 is not directly sharing their repair needs via multicast communication. 2758 The NORM_CMD(REPAIR_ADV) message is multicast to the receiver set by the 2759 sender. The payload portion of this message has content in the same 2760 format as the NORM_NACK receiver message payload. Receivers are then 2761 able to perform feedback suppression in the same manner as with 2762 NORM_NACK messages directly received from other receivers. Note the 2763 sender does not merely retransmit NACK content it receives, but instead 2764 transmits a representation of its aggregated repair state. The 2765 transmission of NORM_CMD(REPAIR_ADV) messages are subject to the sender 2766 transmit rate limit and NormSegmentSize limitation. When the 2767 NORM_CMD(REPAIR_ADV) message is of maximum size, receivers SHALL 2768 consider the maximum ordinal transmission position value embedded in the 2769 message as the senders "current" transmission position and implicitly 2770 suppress requests for ordinally higher repair. For congestion control 2771 operation, the sender may also need to provide information so that 2772 dynamic congestion control feedback can be suppressed as needed among 2773 receivers. This document specifies the NORM-CC Feedback Header 2774 Extension that is applied for baseline NORM-CC operation. If other 2775 congestion control mechanisms are used within a NORM implementation, 2776 other header extensions may be defined. Whatever content format is used 2777 for this purpose should ensure that maximum possible suppression state 2778 is conveyed to the receiver set. 2780 5.5. Additional Protocol Mechanisms 2782 In addition to the principal function of data content transmission and 2783 repair, there are some other protocol mechanisms that help NORM to adapt 2784 to network conditions and play fairly with other coexistent protocols. 2786 5.5.1. Greatest Round-trip Time Collection 2788 For NORM receivers to appropriately scale backoff timeouts and the 2789 senders to use proper corresponding timeouts, the participants must 2790 agree on a common timeout basis. Each NORM sender monitors the round- 2791 trip time of active receivers and determines the group greatest round- 2792 trip time (GRTT). The sender advertises this GRTT estimate in every 2793 message it transmits so that receivers have this value available for 2794 scaling their timers. To measure the current GRTT, the sender 2795 periodically sends NORM_CMD(CC) messages that contain a locally 2796 generated timestamp. Receivers are expected to record this timestamp 2797 along with the time the NORM_CMD(CC) message is received. Then, when 2798 the receivers generate feedback messages to the sender, an adjusted 2799 version of the sender timestamp is embedded in the feedback message 2800 (NORM_NACK or NORM_ACK). The adjustment adds the amount of time the 2801 receiver held the timestamp before generating its response. Upon 2802 receipt of this adjusted timestamp, the sender is able to calculate the 2803 round-trip time to that receiver. 2805 The round-trip time for each receiver is fed into an algorithm that 2806 weights and smoothes the values for a conservative estimate of the GRTT. 2807 The algorithm and methodology are described in the NORM Building Block 2808 document [4] in the section entitled "One-to-Many Sender GRTT 2809 Measurement". A conservative estimate helps feedback suppression at a 2810 small cost in overall protocol repair delay. The sender's current 2811 estimate of GRTT is advertised in the "grtt" field found in all NORM 2812 sender messages. The advertised GRTT is also limited to a minimum of 2813 the nominal inter-packet transmission time given the sender's current 2814 transmission rate and system clock granularity. The reason for this 2815 additional limit is to keep the receiver somewhat "event driven" by 2816 making sure the sender has had adequate time to generate any response to 2817 repair requests from receivers given transmit rate limitations due to 2818 congestion control or configuration. 2820 When the NORM-CC Rate header extension is present in NORM_CMD(CC) 2821 messages, the receivers respond to NORM_CMD(CC) messages as described in 2822 Section 5.5.2, "NORM Congestion Control Operation". The NORM_CMD(CC) 2823 messages are periodically generated by the sender as described for 2824 congestion control operation. This provides for proactive, but 2825 controlled, feedback from the group in the form of NORM_ACK messages. 2827 provides for GRTT feedback even if no NORM_NACK messages are being sent. 2828 If operating without congestion control in a closed network, the 2829 NORM_CMD(CC) messages may be sent periodically without the NORM-CC Rate 2830 header extension. In this case, receivers will only provide GRTT 2831 measurement feedback when NORM_NACK messages are generated since no 2832 NORM_ACK messages are generated. In this case, the NORM_CMD(CC) 2833 messages may be sent less frequently, perhaps as little as once per 2834 minute, to conserve network capacity. Note that the NORM-CC Rate header 2835 extension may also be used proactively solicit RTT feedback from the 2836 receiver group per congestion control operation even though the sender 2837 may not be conducting congestion control rate adjustment. NORM 2838 operation without congestion control should be considered only in closed 2839 networks. 2841 5.5.2. NORM Congestion Control Operation 2843 This section describes baseline congestion control operation for the 2844 NORM protocol (NORM-CC). The supporting NORM message formats and 2845 approach described here are an adaptation of the equation-based TCP- 2846 Friendly Multicast Congestion Control (TFMCC) approach described in 2847 [19]. This congestion control scheme is REQUIRED for operation within 2848 the general Internet unless the NORM implementation is adapted to use 2849 another IETF-sanctioned reliable multicast congestion control mechanism 2850 (e.g. PGMCC [20]). With this TFMCC-based approach, the transmissions 2851 of NORM senders are controlled in a rate-based manner as opposed to 2852 window-based congestion control algorithms as in TCP. However, it is 2853 possible that the NORM protocol message set may alternatively be used to 2854 support a window-based multicast congestion control scheme such as 2855 PGMCC. The details of that alternative may be described separately or 2856 in a future revision of this document. In either case (rate-based TFMCC 2857 or window-based PGMCC), successful control of sender transmission 2858 depends upon collection of sender-to-receiver packet loss estimates and 2859 RTTs to identify the congestion control bottleneck path(s) within the 2860 multicast topology and adjust the sender rate accordingly. The receiver 2861 with loss and RTT estimates that correspond to the lowest result 2862 transmission rate is identified as the "current limiting receiver" 2863 (CLR). 2865 As described in [21], a steady-state sender transmission rate, to be 2866 "friendly" with competing TCP flows can be calculated as: 2868 S 2869 Rsender = --------------------------------------------------------------- 2870 tRTT * (sqrt((2/3)*p) + 12 * sqrt((3/8)*p) * p * (1 + 32*(p^2))) 2872 where 2874 S = Nominal transmitted packet size. (In NORM, the "nominal" 2875 packet size can be determined by the sender as an 2876 exponentially weighted moving average (EWMA) of transmitted 2877 packet sizes to account for variable message sizes). 2879 tRTT = The RTT estimate of the current "current limiting receiver" 2880 (CLR). 2882 p = The loss event fraction of the CLR. 2884 To support congestion control feedback collection and operation, the 2885 NORM sender periodically transmits NORM_CMD(CC) command messages. 2886 NORM_CMD(CC) messages are multiplexed with NORM data and repair 2887 transmissions and serve several purposes: 2889 1) Stimulate explicit feedback from the general receiver set to 2890 collect congestion control information. 2892 2) Communicate state to the receiver set on the sender's 2893 current congestion control status including details of the 2894 CLR. 2896 3) Initiate rapid (immediate) feedback from the CLR in order to 2897 closely track the dynamics of congestion control for that 2898 current "worst path" in the group multicast topology. 2900 The format of the NORM_CMD(CC) message is describe in Section 4.2.3 of 2901 this document. The NORM_CMD(CC) message contains information to allow 2902 measurement of RTTs, to inform the group of the congestion control CLR, 2903 and to provide feedback of individual RTT measurements to the receivers 2904 in the group. The NORM_CMD(CC) also provides for exciting feedback from 2905 OPTIONAL "potential limiting receiver" (PLR) nodes that may be 2906 determined administratively or possibly algorithmically based on 2907 congestion control feedback. PLR nodes are receivers that have been 2908 identified to have potential for (perhaps soon) becoming the CLR and 2909 thus immediate, up-to-date feedback is beneficial for congestion control 2910 performance. The details of PLR selection are not discussed in this 2911 document. 2913 5.5.2.1. NORM_CMD(CC) Transmission 2915 The NORM_CMD(CC) message is transmitted periodically by the sender along 2916 with its normal data transmission. Note that the repeated transmission 2917 of NORM_CMD(CC) messages may be initiated some time before transmission 2918 of user data content at session startup. This may be done to collect 2919 some estimation of the current state of the multicast topology with 2920 respect to group and individual RTT and congestion control state. 2922 A NORM_CMD(CC) message is immediately transmitted at sender startup. 2923 The interval of subsequent NORM_CMD(CC) message transmission is 2924 determined as follows: 2926 1) By default, the interval is set according to the current 2927 sender GRTT estimate. A startup GRTT of 0.5 seconds is 2928 recommended when no feedback has yet been received from the 2929 group. 2931 2) If a CLR has been identified (based on previous receiver 2932 feedback), the interval is the RTT between the sender and 2933 CLR. 2935 3) Additionally, if the interval of nominal data message 2936 transmission is greater than the GRTT or RTT_clr interval, 2937 the NORM_CMD(CC) interval is set to this greater value. 2938 This ensures that the transmission of this control message 2939 is not done to the exclusion of user data transmission. 2941 The NORM_CMD(CC) "cc_sequence" field is incremented with each 2942 transmission of a NORM_CMD(CC) command. The greatest "cc_sequence" 2943 recently received by receivers is included in their feedback to the 2944 sender. This allows the sender to determine the "age" of feedback to 2945 assist in congestion avoidance. 2947 The NORM-CC Rate Header Extension is applied to the NORM_CMD(CC) message 2948 and the sender advertises its current transmission rate in the 2949 "send_rate" field. The rate information is used by receivers to 2950 initialize loss estimation during congestion control startup or restart. 2952 The "cc_node_list" contains a list of entries identifying receivers and 2953 their current congestion control state (status "flags", "rtt" and "loss" 2954 estimates). The list may be empty if the sender has not yet received 2955 any feedback from the group. If the sender has received feedback, the 2956 list will minimally contain an entry identifying the CLR. A 2957 NORM_FLAG_CC_CLR flag value is provided for the "cc_flags" field to 2958 identify the CLR entry. It is RECOMMENDED that the CLR entry be the 2959 first in the list for implementation efficiency. Additional entries in 2960 the list are used to provide sender-measured individual RTT estimates to 2961 receivers in the group. The number of additional entries in this list 2962 is dependent upon the percentage of control traffic the sender 2963 application is willing to send with respect to user data message 2964 transmissions. More entries in the list may allow the sender to be more 2965 responsive to congestion control dynamics. The length of the list may 2966 be dynamically determined according to the current transmission rate and 2967 scheduling of NORM_CMD(CC) messages. The maximum length of the list 2968 corresponds to the sender's NormSegmentSize parameter for the session. 2969 The inclusion of additional entries in the list based on receiver 2970 feedback are prioritized with following rules: 2972 1) Receivers that have not yet been provided RTT feedback get 2973 first priority. Of these, those with the greatest loss 2974 fraction receive precedence for list inclusion. 2976 2) Secondly, receivers that have previously been provided RTT 2977 are included with receivers yielding the lowest calculated 2978 congestion rate getting precedence. 2980 There are "cc_flag" values in addition to NORM_FLAG_CC_CLR that are used 2981 for other congestion control functions. The NORM_FLAG_CC_PLR flag value 2982 is used to mark additional receivers from that the sender would like to 2983 have immediate, non-suppressed feedback. These may be receivers that 2984 the sender algorithmically identified as potential future CLRs or that 2985 have been pre-configured as potential congestion control points in the 2986 network. The NORM_FLAG_CC_RTT indicates the validity of the "cc_rtt" 2987 field for the associated receiver node. Normally, this flag will be set 2988 since the receivers in the list will typically be receivers from which 2989 the sender has received feedback. However, in the case that the NORM 2990 sender has been pre-configured with a set of PLR nodes, feedback from 2991 those receivers may not yet have been collected and thus the "cc_rtt" 2992 and "cc_rate" fields do not contain valid values when this flag is not 2993 set. 2995 5.5.2.2. NORM_CMD(CC) Feedback Response 2997 Receivers explicitly respond to NORM_CMD(CC) messages in the form of a 2998 NORM_ACK(RTT) message. The goal of the congestion control feedback is 2999 to determine the receivers with the lowest congestion control rates. 3000 Receivers that are marked as CLR or PLR nodes in the NORM_CMD(CC) 3001 "cc_node_list" immediately provide feedback in the form of a NORM_ACK to 3002 this message. When a NORM_CMD(CC) is received, non-CLR or non-PLR nodes 3003 initiate random feedback backoff timeouts similar to that used when the 3004 receiver initiates a repair cycle (see Section 5.3 ) in response to 3005 detection of data loss. The backoff timeout for the congestion control 3006 response is generated as follows: 3008 T_backoff = RandomBackoff(K*GRTTsender, GSIZEsender) 3010 The "RandomBackoff()" algorithm provides a truncated exponentially 3011 distributed random number and is described in the NORM Building Block 3012 document [4]. The same backoff factor K = Ksender MAY be used as with 3013 NORM_NACK suppression. However, in cases where the application 3014 purposefully specifies a very small Ksender backoff factor to minimize 3015 the NACK repair process latency (trading off group size scalability), it 3016 may still be desirable to maintain a larger backoff factor for 3017 congestion control feedback, since there may often be a larger volume of 3018 congestion control feedback than NACKs in many cases and congestion 3019 control feedback latency may be tolerable where reliable delivery 3020 latency is not. As previously noted, a backoff factor value of K = 4 is 3021 generally recommended for ASM operation and K = 6 for SSM operation. A 3022 receiver SHALL cancel the backoff timeout and thus its pending 3023 transmission of a NORM_ACK(RTT) message under the following conditions: 3025 1) The receiver generates another feedback message (NORM_NACK 3026 or other NORM_ACK) before the congestion control feedback 3027 timeout expires, 3029 2) A NORM_CMD(CC) or other receiver feedback with an ordinally 3030 greater "cc_sequence" field value is received before the 3031 congestion control feedback timeout expires (This is similar 3032 to the TFMCC feedback round number), 3034 3) When the T_backoff is greater than 1*GRTT. This prevents 3035 NACK implosion in the event of sender or network failure, 3037 4) "Suppressing" congestion control feedback is heard from 3038 another receiver (in a NORM_ACK or NORM_NACK) or via a 3039 NORM_CMD(REPAIR_ADV) message from the sender. The local 3040 receiver's feedback is "suppressed" if the rate of the 3041 competing feedback (Rfb) is sufficiently close to or less 3042 than the local receiver's calculated rate (Rcalc). The 3043 local receiver's feedback is canceled when: 3045 Rcalc > (0.9 * Rfb) 3047 Also note receivers that have not yet received an RTT 3048 measurement from the sender are suppressed only by other 3049 receivers that have not yet measured RTT. Additionally, 3050 receivers whose RTT estimate has "aged" considerably (i.e. 3051 they haven't been included in the NORM_CMD(CC) 3052 "cc_node_list" in a long time) may wish to compete as a 3053 receiver with no prior RTT measurement after some expiration 3054 period. 3056 When the backoff timer expires, the receiver SHALL generate a 3057 NORM_ACK(RTT) message to provide feedback to the sender and group. This 3058 message may be multicast to the group for most effective suppression in 3059 ASM topologies or unicast to the sender depending upon how the NORM 3060 protocol is deployed and configured. 3062 Whenever any feedback is generated (including this NORM_ACK(RTT) 3063 message), receivers include an adjusted version of the sender timestamp 3064 from the most recently received NORM_CMD(CC) message and the 3065 "cc_sequence" value from that command in the applicable NORM_ACK or 3066 NORM_NACK message fields. For NORM-CC operation, any generated feedback 3067 message SHALL also contain the NORM-CC Feedback header extension. The 3068 receiver provides its current "cc_rate" estimate, "cc_loss" estimate, 3069 "cc_rtt" if known, and any applicable "cc_flags" via this header 3070 extension. 3072 During slow start (when the receiver has not yet detected loss from the 3073 sender), the receiver uses a value equal to two times its measured rate 3074 from the sender in the "cc_rate" field. For steady-state congestion 3075 control operation, the receiver "cc_rate" value is from the equation- 3076 based value using its current loss event estimate and sender<->receiver 3077 information. (The GRTT is used when the receiver has not yet measured 3078 its individual RTT). 3080 The "cc_loss" field value reflects the receiver's current loss event 3081 estimate with respect to the sender in question. 3083 When the receiver has a valid individual RTT measurement, it SHALL 3084 include this value in the "cc_rtt" field. The NORM_FLAG_CC_RTT MUST be 3085 set when the "cc_rtt" field is valid. 3087 After a congestion control feedback message is generated or when the 3088 feedback is suppressed, a non-CLR receiver begins a "holdoff" timeout 3089 period during which it will restrain itself from providing congestion 3090 control feedback, even if NORM_CMD(CC) messages are received from the 3091 sender (unless the receive becomes marked as a CLR or PLR node). The 3092 value of this holdoff timeout (T_ccHoldoff) period is: 3094 T_ccHoldoff = (K*GRTT) 3096 Thus, non-CLR receivers are constrained to providing explicit congestion 3097 control feedback once per K*GRTT intervals. Note, however, that as the 3098 session progresses, different receivers will be responding to different 3099 NORM_CMD(CC) messages and there will be relatively continuous feedback 3100 of congestion control information while the sender is active. 3102 5.5.2.3. Congestion Control Rate Adjustment 3104 During steady-state operation, the sender will directly adjust its 3105 transmission rate to the rate indicated by the feedback from its 3106 currently selected CLR. As noted in [19], the estimation of parameters 3107 (loss and RTT) for the CLR will generally constrain the rate changes 3108 possible within acceptable bounds. For rate increases, the sender SHALL 3109 observe a maximum rate of increase of one packet per RTT at all times 3110 during steady-state operation. 3112 The sender processes congestion control feedback from the receivers and 3113 selects the CLR based on the lowest rate receiver. Receiver rates are 3114 either determined directly from the slow start "cc_rate" provided by the 3115 receiver in the NORM-CC Feedback header extension or by performing the 3116 equation-based calculation using individual RTT and loss estimates 3117 ("cc_loss") as feedback is received. 3119 The sender can calculate a current RTT for a receiver (RTT_rcvrNew) 3120 using the "grtt_response" timestamp included in feedback messages. When 3121 the "cc_rtt" value in a response is not valid, the sender simply uses 3122 this RTT_rcvrNew value as the receiver's current RTT (RTT_rcvr). For 3123 non-CLR and non-PLR receivers, the sender can use the "cc_rtt" value 3124 provided in the NORM-CC Feedback header extension as the receiver's 3125 previous RTT measurement (RTT_rcvrPrev) to smooth according to: 3127 RTT_rcvr = 0.5 * RTT_rcvrPrev + 0.5 * RTT_rcvrNew 3129 For CLR receivers where feedback is received more regularly, the sender 3130 SHOULD maintain a more smoothed RTT estimate upon new feedback from the 3131 CLR where: 3133 RTT_clr = 0.9 * RTT_clr + 0.1 * RTT_clrNew 3135 "RTT_clrNew" is the new RTT calculated from the timestamp in the 3136 feedback message received from the CLR. The RTT_clr is initialized to 3137 RTT_clrNew on the first feedback message received. Note that the same 3138 procedure is observed by the sender for PLR receivers and that if a PLR 3139 is "promoted" to CLR status, the smoothed estimate can be continued. 3141 There are some additional periods besides steady-state operation that 3142 need to be considered in NORM-CC operation. These periods are: 3144 1) during session startup, 3146 2) when no feedback is received from the CLR, and 3148 3) when the sender has a break in data transmission. 3150 During session startup, the congestion control operation SHALL observe a 3151 "slow start" procedure to quickly approach its fair bandwidth share. An 3152 initial sender startup rate is assumed where: 3154 Rinitial = MIN(NormSegmentSize / GRTT, NormSegmentSize) bytes/second. 3156 The rate is increased only when feedback is received from the receiver 3157 set. The "slow start" phase proceeds until any receiver provides 3158 feedback indicating that loss has occurred. Rate increase during slow 3159 start is applied as: 3161 Rnew = Rrecv_min 3163 where "Rrecv_min" is the minimum reported receiver rate in the "cc_rate" 3164 field of congestion control feedback messages received from the group. 3165 Note that during "slow start", receivers use two times their measured 3166 rate from the sender in the "cc_rate" field of their feedback. Rate 3167 increase adjustment is limited to once per GRTT during slow start. 3169 If the CLR or any receiver intends to leave the group, it will set the 3170 NORM_FLAG_CC_LEAVE in its congestion control feedback message as an 3171 indication that the sender should not select it as the CLR. When the 3172 CLR changes to a lower rate receiver, the sender should immediately 3173 adjust to the new lower rate. The sender is limited to increasing its 3174 rate at one additional packet per RTT towards any new, higher CLR rate. 3176 The sender should also track the "age" of the feedback it has received 3177 from the CLR by comparing its current "cc_sequence" value (Seq_sender) 3178 to the last "cc_sequence" value received from the CLR (Seq_clr). As the 3179 "age" of the CLR feedback increases with no new feedback, the sender 3180 SHALL begin reducing its rate once per RTT_clr as a congestion avoidance 3181 The following algorithm is used to determine the decrease in sender 3182 rate (Rsender bytes/sec) as the CLR feedback, unexpectedly, excessively 3183 ages: 3185 Age = Seq_sender - Seq_clr; 3186 if (Age > 4) Rsender = Rsender * 0.5; 3188 This rate reduction is limited to the lower bound on NORM transmission 3189 rate. After NORM_ROBUST_FACTOR consecutive NORM_CMD(CC) rounds without 3190 any feedback from the CLR, the sender SHOULD assume the CLR has left the 3191 group and pick the receiver with the next lowest rate as the new CLR. 3192 Note this assumes that the sender does not have explicit knowledge that 3193 the CLR intentionally left the group. If no receiver feedback is 3194 received, the sender MAY wish to withhold further transmissions of 3195 NORM_DATA segments and maintain NORM_CMD(CC) transmissions only until 3196 feedback is detected. After such a CLR timeout, the sender will be 3197 transmitting with a minimal rate and should return to slow start as 3198 described here for a break in data transmission. 3200 When the sender has a break in its data transmission, it can continue to 3201 probe the group with NORM_CMD(CC) messages to maintain RTT collection 3202 from the group. This will enable the sender to quickly determine an 3203 appropriate CLR upon data transmission restart. However, the sender 3204 should exponentially reduce its target rate to be used for transmission 3205 restart as time since the break elapses. The target rate SHOULD be 3206 recalculated once per RTT_clr as: 3208 Rsender = Rsender * 0.5; 3210 If the minimum NORM rate is reached, the sender should set the 3211 NORM_FLAG_START flag in its NORM_CMD(CC) messages upon restart and the 3212 group should observer "slow start" congestion control procedures until 3213 any receiver experiences a new loss event. 3215 5.5.3. NORM Positive Acknowledgment Procedure 3217 NORM provides options for the source application to request positive 3218 acknowledgment (ACK) of NORM_CMD(FLUSH) and NORM_CMD(ACK_REQ) messages 3219 from members of the group. There are some specific acknowledgment 3220 requests defined for the NORM protocol and a range of acknowledgment 3221 request types that are left to be defined by the application. One 3222 predefined acknowledgment type is the NORM_ACK_FLUSH type. This 3223 acknowledgment is used to determine if receivers have achieved 3224 completion of reliable reception up through a specific logical 3225 transmission point with respect to the sender's sequence of 3226 transmission. The NORM_ACK_FLUSH acknowledgment may be used to assist 3227 in application flow control when the sender has information on a portion 3228 of the receiver set. Another predefined acknowledgment type is 3229 NORM_ACK(CC), which is used to explicitly provide congestion control 3230 feedback in response to NORM_CMD(CC) messages transmitted by the sender 3231 for NORM-CC operation. Note the NORM_ACK(CC) response does NOT follow 3232 the positive acknowledgment procedure described here. The 3233 NORM_CMD(ACK_REQ) and NORM_ACK messages contain an "ack_type" field to 3234 the type of acknowledgment requested and provided. A range of 3235 "ack_type" values is provided for application-defined use. While the 3236 application is responsible for initiating the acknowledgment request and 3237 interprets application-defined "ack_type" values, the acknowledgment 3238 procedure SHOULD be conducted within the protocol implementation to take 3239 advantage of timing and transmission scheduling information available to 3240 the NORM transport. 3242 The NORM positive acknowledgment procedure uses polling by the sender to 3243 query the receiver group for response. Note this polling procedure is 3244 not intended to scale to very large receiver groups, but could be used 3245 in large group setting to query a critical subset of the group. Either 3246 the NORM_CMD(ACK_REQ), or when applicable, the NORM_CMD(FLUSH) message 3247 is used for polling and contains a list of NormNodeIds for receivers 3248 that should respond to the command. The list of receivers providing 3249 acknowledgment is determined by the source application with "a priori" 3250 knowledge of participating nodes or via some other application-level 3251 mechanism. 3253 The ACK process is initiated by the sender that generates 3254 NORM_CMD(FLUSH) or NORM_CMD(ACK_REQ) messages in periodic "rounds". For 3255 NORM_ACK_FLUSH requests, the NORM_CMD(FLUSH) contain a 3256 "object_transport_id" and "fec_payload_id" denoting the watermark 3257 transmission point for which acknowledgment is requested. This 3258 watermark transmission point is "echoed" in the corresponding fields of 3259 the NORM_ACK(FLUSH) message sent by the receiver in response. 3260 NORM_CMD(ACK_REQ) messages contain an "ack_id" field which is similarly 3261 "echoed" in response so that the sender may match the response to the 3262 appropriate request. 3264 In response to the NORM_CMD(ACK_REQ), the listed receivers randomly 3265 spread NORM_ACK messages uniformly in time over a window of (1*GRTT). 3266 These NORM_ACK messages are typically unicast to the sender. (Note that 3267 NORM_ACK(CC) messages SHALL be multicast or unicast in the same manner 3268 as NORM_NACK messages). 3270 The ACK process is self-limiting and avoids ACK implosion in that: 3272 1) Only a single NORM_CMD(ACK_REQ) message is generated once per 3273 (2*GRTT), and, 3275 2) The size of the "acking_node_list" of NormNodeIds from which 3276 acknowledgment is requested is limited to a maximum of the 3277 sender NormSegmentSize setting per round of the positive 3278 acknowledgment process. 3280 Because the size of the included list is limited to the sender's 3281 NormSegmentSize setting, multiple NORM_CMD(ACK_REQ) rounds may be 3282 required to achieve responses from all receivers specified. The 3283 content of the attached NormNodeId list will be dynamically updated as 3284 this process progresses and NORM_ACK responses are received from the 3285 specified receiver set. As the sender receives valid responses (i.e. 3287 watermark point or "ack_id") from receivers, it SHALL eliminate those 3288 receivers from the subsequent NORM_CMD(ACK_REQ) message 3289 "acking_node_list" and add in any pending receiver NormNodeIds while 3290 keeping within the NormSegmentSize limitation of the list size. Each 3291 receiver is queried a maximum number of times (NORM_ROBUST_FACTOR, by 3292 default). Receivers not responding within this number of repeated 3293 requests are removed from the payload list to make room for other 3294 potential receivers pending acknowledgment. The transmission of the 3295 NORM_CMD(ACK_REQ) is repeated until no further responses are required or 3296 until the repeat threshold is exceeded for all pending receivers. The 3297 transmission of NORM_CMD(ACK_REQ) or NORM_CMD(FLUSH) messages to conduct 3298 the positive acknowledgment process is multiplexed with ongoing sender 3299 data transmissions. However, the NORM_CMD(FLUSH) positive 3300 acknowledgment process may be interrupted in response to negative 3301 acknowledgment repair requests (NACKs) received from receivers during 3302 the acknowledgment period. The NORM_CMD(FLUSH) positive acknowledgment 3303 process is restarted for receivers pending acknowledgment once any the 3304 repairs have been transmitted. 3306 In the case of NORM_CMD(FLUSH) commands with an attached 3307 "acking_node_list", receivers will not ACK until they have received 3308 complete transmission of all data up to and including the given 3309 watermark transmission point. All receivers SHALL interpret the 3310 watermark point provided in the request NACK for repairs if needed as 3311 for NORM_CMD(FLUSH) commands with no attached "acking_node_list". 3313 5.5.4. Group Size Estimate 3315 NORM sender messages contain a "gsize" field that is a representation of 3316 the group size and is used in scaling random backoff timer ranges. The 3317 use of the group size estimate within the NORM protocol does not require 3318 a precise estimation and works reasonably well if the estimate is within 3319 an order of magnitude of the actual group size. By default, the NORM 3320 sender group size estimate may be administratively configured. Also, 3321 given the expected scalability of the NORM protocol for general use, a 3322 default value of 10,000 is recommended for use as the group size 3323 estimate. 3325 It is possible that group size may be algorithmically approximated from 3326 the volume of congestion control feedback messages which follow the 3327 exponentially weighted random backoff. However, the specification of 3328 such an algorithm is currently beyond the scope of this document. 3330 6. Security Considerations 3332 The same security considerations that apply to the NORM, and FEC 3333 Building Blocks also apply to the NORM protocol. In addition to 3334 vulnerabilities that any IP and IP multicast protocol implementation may 3335 be generally subject to, the NACK-based feedback of NORM may be 3336 exploited by replay attacks which force the NORM sender to unnecessarily 3337 transmit repair information. This MAY be addressed by network layer IP 3338 security implementations that guard against this potential security 3339 exploitation. It is RECOMMENDED that such IP security mechanisms be 3340 when available. Another possible approach is for NORM senders to use 3341 the "sequence" field from the NORM Common Message Header to detect 3342 replay attacks. This can be accomplished if the NORM packets are 3343 cryptographically protected and the sender is willing to maintain state 3344 on receivers which are NACKing. A cache of receiver state may provide 3345 some protection against replay attacks. Note that the "sequence" field 3346 of NORM messages should be incremented with independent values for 3347 different destinations (e.g., group-addressed versus unicast-addressed 3348 messages versus "receiver" messages). Thus, the congestion control loss 3349 estimation function of the "sequence" field can be preserved for sender 3350 messages when receiver messages are unicast to the sender. The NORM 3351 protocol is compatible with the use of the IP security (IPsec) 3352 architecture described in [22] It is important to note that while NORM 3353 does leverage FEC-based repair for scalability, this does not alone 3354 guarantee integrity of received data. Application-level integrity- 3355 checking of data content is highly RECOMMENDED. 3357 7. IANA Considerations 3359 No information in this specification is currently subject to IANA 3360 registration. However, several Header Extensions are defined within 3361 this document. If/when additional Header Extensions are developed, the 3362 first RFC MUST establish an IANA registry for them, with a 3363 "Specification Required" policy [6] and all Header Extensions, 3364 including those in the present document, MUST be registered thereafter. 3365 Additionally, building blocks components used by NORM may introduce 3366 additional IANA considerations. In particular, the FEC Building Block 3367 used by NORM does require IANA registration of the FEC codecs used. The 3368 registration instructions for FEC codecs are provided in [5]. 3370 8. Suggested Use 3372 The present NORM protocol is seen as useful tool for the reliable data 3373 transfer over generic IP multicast services. It is not the intention 3374 of the authors to suggest it is suitable for supporting all envisioned 3375 multicast reliability requirements. NORM provides a simple and flexible 3376 framework for multicast applications with a degree of concern for 3377 network traffic implosion and protocol overhead efficiency. NORM-like 3378 protocols have been successfully demonstrated within the MBone for bulk 3379 data dissemination applications, including weather satellite compressed 3380 imagery updates servicing a large group of receivers and a generic web 3381 content reliable "push" application. 3383 In addition, this framework approach has some design features making it 3384 attractive for bulk transfer in asymmetric and wireless internetwork 3385 applications. NORM is capable of successfully operating independent of 3386 network structure and in environments with high packet loss, delay, and 3387 misordering. Hybrid proactive/reactive FEC-based repairing improve 3388 protocol performance in some multicast scenarios. A sender-only repair 3389 approach often makes additional engineering sense in asymmetric 3390 networks. NORM's unicast feedback capability may be suitable for use in 3391 asymmetric networks or in networks where only unidirectional multicast 3392 routing/delivery service exists. Asymmetric architectures supporting 3393 delivery are likely to make up an important portion of the future 3394 Internet structure (e.g., DBS/cable/PSTN hybrids) and efficient, 3395 reliable bulk data transfer will be an important capability for 3396 servicing large groups of subscribed receivers. 3398 9. Acknowledgments (and these are not Negative) 3400 The authors would like to thank Rick Jones, Vincent Roca, Rod Walsh, 3401 Toni Paila, Michael Luby, and Joerg Widmer for their valuable input and 3402 comments on this document. The authors would also like to thank the RMT 3403 working group chairs, Roger Kermode and Lorenzo Vicisano, for their 3404 support in development of this specification, and Sally Floyd for her 3405 early input into this document. 3407 10. References 3409 10.1. Normative References 3411 [1] R. Kermode, L. Vicisano, "Author Guidelines for Reliable Multicast 3412 Transport (RMT) Building Blocks and Protocol Instantiation documents", 3413 RFC 3269, April 2002. 3415 [2] S. Bradner, "Key words for use in RFCs to Indicate Requirement 3416 Levels", BCP 14, RFC 2119, March 1997. 3418 [3] S. Deering, "Host Extensions for IP Multicasting", STD 5, RFC 1112, 3419 August 1989. 3421 [4] B. Adamson, C. Bormann, M. Handley, and J. Macker, "NACK-Oriented 3422 Reliable Multicast (NORM) Protocol Building Blocks", draft-ietf-rmt-bb- 3423 norm-09, July 2004. 3425 [5] M. Luby, L. Vicisano, J. Gemmell, L. Rizzo, M. Handley, and J. 3426 Crowcroft, "Forward Error Correction (FEC) Building Block", RFC 3452, 3427 December 2002. 3429 [6] T. Narten and H. Alvestrand, "Guidelines for Writing an IANA 3430 Considerations Section in RFCs", BCP 14, RFC 2434, October 1998. 3432 10.2. Informative References 3434 [7] M. Handley, and V. Jacobson, "SDP: Session Description Protocol", 3435 RFC 2327, April 1998. 3437 [8] M. Handley, C. Perkins, and E. Whelan, "Session Announcement 3438 Protocol", RFC 2974, October 2000. 3440 [9] S. Pingali, D. Towsley, J. Kurose, "A Comparison of Sender-Initiated 3441 and Receiver-Initiated Reliable Multicast Protocols", In Proc. INFOCOM, 3442 San Francisco CA, October 1993. 3444 [10] M. Luby, L. Vicisano, J. Gemmell, L. Rizzo, M. Handley, and J. 3445 Crowcroft, "The Use of Forward Error Correction (FEC) in Reliable 3446 Multicast", RFC 3453, December 2002. 3448 [11] J. Macker, B. Adamson, "The Multicast Dissemination Protocol (MDP) 3449 Toolkit", Proc. IEEE MILCOM 99, October 1999. 3451 [12] J. Nonnenmacher and E. Biersack, "Optimal Multicast Feedback", 3452 Proc. IEEE INFOCOMM, p. 964, March/April 1998. 3454 [13] J. Macker, B. Adamson, "Quantitative Prediction of Nack Oriented 3455 Reliable Multicast (NORM) Feedback", Proc. IEEE MILCOM 2002, October 3456 2002. 3458 [14] H.W. Holbrook, "A Channel Model for Multicast", Ph.D. Dissertation, 3459 Stanford University, Department of Computer Science, Stanford, 3460 California, August 2001. 3462 [15] D. Gossink, J. Macker, "Reliable Multicast and Integrated Parity 3463 Retransmission with Channel Estimation", IEEE GLOBECOMM 98', September 3464 1998. 3466 [16] Whetten, B., Vicisano, L., Kermode, R., Handley, M., Floyd S. and 3467 Luby, M., "Reliable Multicast Transport Building Blocks for One-to-Many 3468 Bulk-Data Transfer", RFC 3048, January 2001. 3470 [17] A. Mankin, A. Romanow, S. Bradner, and V. Paxson, "IETF Criteria 3471 for Evaluating Reliable Multicast Transport and Application Protocols", 3472 RFC 2357, June 1998. 3474 [18] H. Schulzrinne, S. Casner, R. Frederick, V. Jacobson, "RTP: A 3475 Transport Protocol for Real-Time Applications", RFC 1889, January 1996. 3477 [19] J. Widmer and M. Handley, "Extending Equation-Based Congestion 3478 Control to Multicast Applications", Proc ACM SIGCOMM 2001, San Diego, 3479 August 2001. 3481 [20] L. Rizzo, "pgmcc: A TCP-Friendly Single-Rate Multicast Congestion 3482 Control Scheme", Proc ACM SIGCOMM 2000, Stockholm, August 2000. 3484 [21] J. Padhye, V. Firoiu, D. Towsley, and J. Kurose, "Modeling TCP 3485 Throughput: A Simple Model and its Empirical Validation", Proc ACM 3486 SIGCOMM 1998. 3488 [22] S. Kent and R. Atkinson, "Security Architecture for the Internet 3489 Protocol", RFC 2401, November 1998. 3491 11. Authors' Addresses 3493 Brian Adamson 3494 adamson@itd.nrl.navy.mil 3495 Naval Research Laboratory 3496 Washington, DC, USA, 20375 3498 Carsten Bormann 3499 cabo@tzi.org 3500 Universitaet Bremen TZI 3501 Postfach 330440 3502 D-28334 Bremen, Germany 3504 Mark Handley 3505 mjh@aciri.org 3506 1947 Center Street, Suite 600 3507 Berkeley, CA 94704 3509 Joe Macker 3510 macker@itd.nrl.navy.mil 3511 Naval Research Laboratory 3512 Washington, DC, USA, 20375 3514 12. Full Copyright Statement 3516 Copyright (C) The Internet Society (2004). This document is subject to 3517 the rights, licenses and restrictions contained in BCP 78 and except as 3518 set forth therein, the authors retain all their rights. 3520 This document and the information contained herein are provided on an 3521 "AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS OR 3522 IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY AND THE INTERNET 3523 ENGINEERING TASK FORCE DISCLAIM ALL WARRANTIES, EXPRESS OR IMPLIED, 3524 INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE 3525 INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED 3526 WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.