idnits 2.17.1 draft-ietf-issll-atm-mapping-02.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- ** Cannot find the required boilerplate sections (Copyright, IPR, etc.) in this document. Expected boilerplate is as follows today (2024-04-23) according to https://trustee.ietf.org/license-info : IETF Trust Legal Provisions of 28-dec-2009, Section 6.a: This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79. IETF Trust Legal Provisions of 28-dec-2009, Section 6.b(i), paragraph 2: Copyright (c) 2024 IETF Trust and the persons identified as the document authors. All rights reserved. IETF Trust Legal Provisions of 28-dec-2009, Section 6.b(i), paragraph 3: This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (https://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- ** Missing expiration date. The document expiration date should appear on the first and last page. ** The document seems to lack a 1id_guidelines paragraph about Internet-Drafts being working documents. ** The document seems to lack a 1id_guidelines paragraph about 6 months document validity -- however, there's a paragraph with a matching beginning. Boilerplate error? ** The document seems to lack a 1id_guidelines paragraph about the list of current Internet-Drafts. ** The document seems to lack a 1id_guidelines paragraph about the list of Shadow Directories. == No 'Intended status' indicated for this document; assuming Proposed Standard Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- ** The document seems to lack an IANA Considerations section. (See Section 2.2 of https://www.ietf.org/id-info/checklist for how to handle the case when there are no actions for IANA.) ** The document seems to lack separate sections for Informative/Normative References. All references will be assumed normative when checking for downward references. ** There are 78 instances of too long lines in the document, the longest one being 7 characters in excess of 72. == There are 1 instance of lines with non-RFC2606-compliant FQDNs in the document. Miscellaneous warnings: ---------------------------------------------------------------------------- == Line 65 has weird spacing: '...Service x...' == Line 95 has weird spacing: '...egories x...' == Line 923 has weird spacing: '...mber of usefu...' -- 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 (26 September 1997) is 9706 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) == Unused Reference: '3' is defined on line 1511, but no explicit reference was found in the text == Unused Reference: '4' is defined on line 1514, but no explicit reference was found in the text == Unused Reference: '5' is defined on line 1517, but no explicit reference was found in the text ** Downref: Normative reference to an Informational RFC: RFC 1633 (ref. '1') -- Possible downref: Non-RFC (?) normative reference: ref. '2' -- Possible downref: Non-RFC (?) normative reference: ref. '3' -- Possible downref: Non-RFC (?) normative reference: ref. '4' -- Possible downref: Non-RFC (?) normative reference: ref. '5' -- Possible downref: Non-RFC (?) normative reference: ref. '6' -- Possible downref: Non-RFC (?) normative reference: ref. '7' -- Possible downref: Non-RFC (?) normative reference: ref. '8' == Outdated reference: A later version (-04) exists of draft-ietf-intserv-ctrl-load-svc-03 -- Possible downref: Non-RFC (?) normative reference: ref. '11' -- Possible downref: Non-RFC (?) normative reference: ref. '12' -- Possible downref: Non-RFC (?) normative reference: ref. '13' -- Possible downref: Non-RFC (?) normative reference: ref. '14' ** Downref: Normative reference to an Informational RFC: RFC 1821 (ref. '15') ** Obsolete normative reference: RFC 1483 (ref. '16') (Obsoleted by RFC 2684) Summary: 12 errors (**), 0 flaws (~~), 9 warnings (==), 13 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 INTERNET-DRAFT Mark W. Garrett, 3 Bellcore 4 Expires 26 September 1997 5 Marty Borden, 6 New Oak Communications 8 26 March 1997 10 Interoperation of Controlled-Load and Guaranteed Services with ATM 11 13 Status of this Memo 15 This document is an Internet-Draft. Internet-Drafts are working 16 documents of the Internet Engineering Task Force (IETF), its areas, 17 and its working groups. Note that other groups may also distribute 18 working documents as Internet-Drafts. 20 Internet-Drafts are draft documents valid for a maximum of six months 21 and may be updated, replaced, or obsoleted by other documents at any 22 time. It is inappropriate to use Internet- Drafts as reference 23 material or to cite them other than as ``work in progress.'' 25 To learn the current status of any Internet-Draft, please check the 26 ``1id-abstracts.txt'' listing contained in the Internet- Drafts 27 Shadow Directories on ftp.is.co.za (Africa), nic.nordu.net (Europe), 28 munnari.oz.au (Pacific Rim), ds.internic.net (US East Coast), or 29 ftp.isi.edu (US West Coast). 31 Abstract 33 Service mappings are an important aspect of effective interoperation 34 between Internet Integrated Services and ATM networks. This document 35 provides guidelines for ATM virtual connection features and 36 parameters to be used in support of the IP integrated services 37 protocols. The specifications include IP Guaranteed Service, 38 Controlled-Load Service and ATM Forum UNI specification, versions 39 3.0, 3.1 and 4.0. 41 These service mappings are intended to facilitate effective end-to- 42 end Quality of Service for IP networks containing ATM subnetworks. 43 We discuss the various features of the IP and ATM protocols, and 44 identify solutions and difficult issues of compatibility and 45 interoperation. 47 Table of Contents 49 0.0 What's New in This Version ......................................... 3 51 1.0 Introduction ....................................................... x 52 1.1 General System Architecture .................................... x 53 1.2 Related Documents .............................................. x 55 2.0 Discussion of ATM Protocol Features ................................ x 56 2.1 Service Category and Bearer Capability ......................... x 57 2.1.1 Service Categories for Guaranteed Service ................ x 58 2.1.2 Service Categories for Controlled Load ................... x 59 2.1.3 Service Categories for Best Effort ....................... x 60 2.2 Cell Loss Priority Bit, Tagging and Conformance Definitions .... x 61 2.3 ATM Adaptation Layer ........................................... x 62 2.4 Broadband Low Layer Information ................................ x 63 2.5 Traffic Descriptors ............................................ x 64 2.5.1 Translating Traffic Descriptors for Guaranteed Service ... x 65 2.5.2 Translating Traffic Descriptors for Controlled Load Service x 66 2.5.3 Translating Traffic Descriptors for Best Effort Service .... x 67 2.6 QoS Classes and Parameters ..................................... x 68 2.7 Additional Parameters -- Frame Discard Mode .................... x 70 3.0 Discussion of IP-IS Protocol Features .............................. x 71 3.1 Handling of Excess Traffic ..................................... x 72 3.2 Use of AdSpec in Guaranteed Service with ATM ................... x 74 4.0 Discussion of Miscellaneous Items .................................. x 75 4.1 Units Conversion ............................................... x 77 5.0 Summary of ATM VC Setup Parameters for Guaranteed Service .......... x 78 5.1 Encoding GS Using Real-Time VBR ................................ x 79 5.2 Encoding GS Using CBR .......................................... x 80 5.3 Encoding GS Using Non-Real-Time VBR ............................ x 81 5.4 Encoding GS Using ABR .......................................... x 82 5.5 Encoding GS Using UBR .......................................... x 83 5.6 Encoding GS Using UNI 3.0 and UNI 3.1. ......................... x 85 6.0 Summary of ATM VC Setup Parameters for Controlled Load Service ..... x 86 6.1 Encoding CLS Using ABR ......................................... x 87 6.2 Encoding CLS Using Non-Real-Time VBR ........................... x 88 6.3 Encoding CLS Using Real-Time VBR ............................... x 89 6.4 Encoding CLS Using CBR ......................................... x 90 6.5 Encoding CLS Using UBR ......................................... x 91 6.6 Encoding CLS Using UNI 3.0 and UNI 3.1. ........................ x 93 7.0 Summary of ATM VC Setup Parameters for Best Effort Service ......... x 94 7.1 Encoding Best Effort Service Using UBR ......................... x 95 7.2 Encoding Best Effort Service Using Other ATM Service Categories x 97 8.0 Acknowledgements ................................................... x 99 Appendix 1 Abbreviations .............................................. x 100 REFERENCES ............................................................. x 101 AUTHORS' ADDRESSES ..................................................... x 103 0.0 What's New in This Version 105 Corrections to VC setup parameter tables. 107 Deleted specific QoS parameter values in tables. 109 Section 3.1 on handling of excess traffic. 111 1.0 Introduction 113 We consider the problem of providing IP Integrated Services [1] with 114 an ATM subnetwork. This document is intended to be consistent with 115 the rsvp protocol [2] for IP-level resource reservation (although it 116 is, strictly speaking, independent of rsvp, since GS and CLS services 117 can be supported through other mechanisms). In the ATM network, we 118 consider ATM Forum UNI Signaling, versions 3.0, 3.1 and 4.0 [3, 4, 119 5]. The latter uses the more complete service model of The ATM 120 Forum's TM 4.0 specification [6, 7]. 122 This is a complex problem with many facets. In this document, we 123 focus on the service types, parameters and signalling elements needed 124 for service interoperation. The resulting service mappings can be 125 used to provide effective end-to-end Quality of Service (QoS) for IP 126 traffic that traverses ATM networks. 128 The IP services considered are Guaranteed Service (GS) [8] and 129 Controlled Load Service (CLS) [9]. We also treat the default Best 130 Effort Service (BE) in parallel with these. Our recommendations for 131 BE are intended to be consistent with RFC 1755 [10], and its revision 132 (in progress) [11], which defines how ATM VCs can be used in support 133 of normal BE IP service. The ATM services we consider are: 135 CBR Constant Bit Rate 136 rtVBR Real-time Variable Bit Rate 137 nrtVBR Non-real-time Variable Bit Rate 138 UBR Unspecified Bit Rate 139 ABR Available Bit Rate 141 (Note, Appendix 1 provides definitions for all abbreviations.) In 142 the case of UNI 3.0 and 3.1 signaling, where these service are not 143 all clearly distinguishable, we identify the appropriate available 144 services. 146 The service mappings which follow most naturally from the service 147 definitions are as follows: 149 Guaranteed Service -> CBR or rtVBR 150 Controlled Load -> nrtVBR or ABR (with a minimum cell rate) 151 Best Effort -> UBR or ABR 153 For completeness we provide detailed mappings for all service 154 combinations and identify how each meets or fails to meet the 155 requirements of the higher level IP services. The reason for not 156 restricting mappings to the most obvious or natural ones is that we 157 cannot assume now that these services will always be ubiquitously 158 available. A number of details, such as treatment of packets in 159 excess of the flow traffic descriptor, make service mapping a 160 complicated subject, which cannot be expressed briefly and accurately 161 at the same time. 163 The remainder of this introduction provides a general discussion of 164 the system configuration and other assumptions. Section 2 considers 165 the relevant ATM protocol elements and their effects as related to 166 Guaranteed, Controlled Load and Best Effort services (the latter 167 being the default "service"). Section 3 discusses a number of 168 important features of the IP services and how they can be handled on 169 an ATM subnetwork. Section 4 addresses a few miscellaneous problems 170 which are neither distinctly IP nor ATM. Section 5 gives detailed VC 171 setup parameters for Guaranteed Service, and considers the effect of 172 using each of the ATM service categories. Section 6 provides a 173 similar treatment for Controlled Load Service. Section 7 considers 174 Best Effort service. 176 This document is only a part of the total solution to providing the 177 interworking of IP integrated services with ATM subnetworks. The 178 important issue of VC management, including flow aggregation, is 179 considered in [12]. We do not consider how routing -- QoS sensitive 180 or not -- interacts with the use of VCs, especially in the case of 181 multicast (or point-to-multipoint) flows. We expect that a 182 considerable degree of implementation latitude will exist, even 183 within the guidelines presented here. Many aspects of interworking 184 between IP and ATM will depend on economic factors, and will not be 185 subject to standardization. 187 1.1 General System Architecture 189 We assume that the reader has a general working knowledge of IP, rsvp 190 and ATM protocols. The network architecture we consider is 191 illustrated in Figure 1, below. An IP-attached host may send unicast 192 datagrams to another host, or may use an IP multicast address to send 193 packets to all of the hosts which have "joined" the multicast "tree". 194 In either case, a destination host may then use RSVP to establish 195 resource reservation in routers along the internet path for the data 196 flow. 198 An ATM network lies in the path (chosen by the IP routing), and 199 consists of one or many ATM switches. It uses VCs to provide both 200 resources and QoS within the ATM cloud. These connections are set 201 up, added to (in the case of multipoint trees), torn down, and 202 controlled by the edge devices, which act as both IP routers and ATM 203 interfaces, capable of initiating and managing VCs across the ATM 204 user-to-network (UNI) interface. The edge devices are assumed to be 205 fully functional in both the IP int-serv/RSVP protocols and the ATM 206 UNI protocols, as well as translating between them. 208 ATM Cloud 209 ------------------ 210 H ----\ ( ) /------- H 211 H ---- R -- R -- E --( ATM Sw -- ATM Sw ) -- E -- R -- R -- H 212 H ----/ | ( ) \ 213 | ------------------ \------ H 214 H ----------R 216 Figure 1: Network Architecture with hosts (H), 217 Routers (R) and Edge Devices (E). 219 The edge devices may be considered part of the IP internet or part of 220 the ATM cloud, or both. This is not an issue since they must provide 221 capabilities of both environments. The edge devices have normal RSVP 222 capability to process RSVP messages, reserve resources, and maintain 223 soft state (in the control path), and to classify and schedule 224 packets (in the data path). They also have the normal ATM 225 capabilities to initiate connections by signaling, and to accept or 226 refuse connections signaled to them. They police and schedule cells 227 going into the ATM cloud. An IP-level reservation (RESV message) 228 triggers the edge device to translate the RSVP service requirements 229 into ATM VC (UNI) semantics. 231 A range of VC management policies are possible, which determine 232 whether a flow should initiate a new VC or join an existing one. VCs 233 are managed according to a combination of standards and local policy 234 rules, which are specific to either the implementation (equipment) or 235 the operator (network service provider). Point-to-multipoint 236 connections within the ATM cloud can be used to support general IP 237 multicast flows. In ATM, a point to multipoint connection can be 238 controlled by the source (or root) node, or a leaf initiated join 239 (LIJ) feature in ATM may be used. Note, the topic of VC management 240 and mapping of flows onto VCs is considered at length in another 241 issll working group draft [12]. 243 Figure 2 shows the functions of an edge device, summarizing the work 244 not part of IP or ATM abstractly as an InterWorking Function (IWF), 245 and segregating the control and data planes. (Note: for expositional 246 convenience, policy control and other control functions are included 247 as part of the admission control in the diagram.) 249 IP ATM 250 ____________________ 251 | IWF | 252 | | 253 admission <--> | service mapping | <--> ATM 254 control | VC management | signalling & 255 | address resolution | admission 256 |....................| control 257 | | 258 classification/ |ATM Adaptation Layer| cell 259 policing & <--> | Segmentation and | <--> scheduling/ 260 scheduling | Reassembly | shaping 261 | Buffering | 262 ____________________ 264 Figure 2: Edge Device Functions showing the IWF 266 In the logical view of Figure 2, some functions, such as scheduling, 267 are shown separately, since these functions are required of both the 268 IP and ATM sides. However it may be possible in an integrated 269 implementation to combine such functions. 271 It is not possible to completely separate the service mapping and VC 272 management functions. Several illustrative examples come to mind: 273 (i) Multiple integrated-services flows may be aggregated to use one 274 point-to-multipoint VC. In this case, we assume the IP flows are of 275 the same service type and their parameters have been merged 276 appropriately. (ii) The VC management function may choose to 277 allocate extra resources in anticipation of further reservations or 278 based on an empiric of changing TSpecs. (iii) There must exist a 279 path for best effort flows and for sending the rsvp control messages. 280 How this interacts with the establishment of VCs for QoS traffic may 281 alter the characteristics required of those VCs. See [12] for 282 further details on VC management. 284 Therefore, in discussing the service-mapping problem, we will assume 285 that the VC management function of the IWF can always express its 286 result in terms of an IP-level service with some QoS and TSpec. The 287 service mapping algorithm, which is the subject of this document, can 288 then identify the appropriate VC parameters, whether the resulting 289 action is initiation of a new VC, the addition/deletion of a leaf to 290 an existing multipoint tree, or the modification of an existing VC to 291 one of another description. 293 1.2 Related Documents 295 Earlier ATM Forum documents were called UNI 3.0 and UNI 3.1. The 3.1 296 release was used to correct errors and fix alignment with the ITU. 297 Unfortunately UNI 3.0 and 3.1 are incompatible. However this is in 298 terms of actual codepoints, not semantics. Therefore, descriptions 299 of parameter values can generally be used for both. 301 After 3.1, the ATM Forum decided to release documents separately for 302 each technical working group. The Traffic Management and Signalling 303 4.0 documents are available publically at ftp.atmforum.com/pub. We 304 refer to the combination of traffic management and signalling as 305 TM/UNI 4.0, although specific references may be made to the TM 4.0 306 specification or the UNI SIG 4.0 specification. 308 Within the IETF area, related material includes the work of the rsvp 309 [2], int-serv [1, 8, 9, 13, 14] and ion working groups [10, 11] of 310 the IETF. Rsvp defines the resource reservation protocol (which is 311 analogous to signaling in ATM). Int-serv defines the behavior and 312 semantics of particular services (analogous e.g., to the Traffic 313 Management working group in the ATM Forum). Ion defines interworking 314 of IP and ATM for traditional Best Effort service, and covers all 315 issues related to routing and addressing. 317 A large number of ATM signaling details are covered in RFC 1755 [10], 318 e.g., differences between UNI 3.0 and UNI 3.1, encapsulation, frame- 319 relay interworking, etc. These considerations generally extend to IP 320 over ATM with QoS as well. Any description given in this document of 321 IP Best Effort service (i.e. the default behavior) over ATM is 322 intended to be consistent with RFC 1755 and it's extension for UNI 323 4.0 [11], and those documents are to be considered definitive. In 324 some instances with non-best-effort services, certain IP/ATM features 325 will diverge from the following RFC 1755. The authors have attempted 326 to note such differences explicitly. (For example, best effort VCs 327 are taken down on timeout by either edge device, while QoS VCs are 328 only removed by the upstream edge device when the corresponding rsvp 329 reservation is deleted.) 331 RFC 1821 [15], represents an early discussions of issues involved 332 with interoperating IP and ATM protocols for integrated services and 333 QoS. 335 2.0 Discussion of ATM Protocol Features 337 In this section, we discuss each of the items that must be specified 338 in the setup of an ATM VC. For each of these we discuss which 339 specified items and values may be most appropriate for each of the 340 three integrated services. 342 The ATM Call Setup is sent by the edge device to the ATM network to 343 establish end-to-end (ATM) service. This setup contains the 344 following information. 346 Service Category/Broadband Bearer Capability 347 AAL Parameters 348 Broadband Low Layer Information 349 Calling and Called Party Addressing Information 350 Traffic Descriptors 351 QoS Parameters 352 Additional Parameters of TM/UNI 4.0 354 We will discuss each of these, except addressing information, as they 355 relate to the translation of GS and CLS to ATM services. Following 356 the discussion of the service categories, we discuss the tagging and 357 conformance definitions for IP and ATM, since the policing method is 358 implicit in the call setup. We then continue with mappings of the 359 other parameters and information elements. 361 2.1 Service Category and Bearer Capability 363 The highest level of abstraction distinguishing features of ATM VCs 364 is in the service category or bearer capability. Service categories 365 were introduced in TM/UNI 4.0; previously the bearer capability was 366 used to discriminate at this level. 368 In each version of the ATM specifications, these indicate the general 369 properties required of a VC: whether there is a real-time delay 370 constraint, whether the traffic is constant or variable rate, the 371 applicable traffic and QoS description parameters and (implicitly) 372 the complexity of some supporting switch mechanisms. 374 For UNI 3.0 and UNI 3.1, there are only two distinct options for 375 bearer capabilities (in our context): 377 BCOB-A: constant rate, timing required, unicast/multipoint; 378 BCOB-C: variable rate, timing not required, unicast/multipoint. 380 There is a third capability, BCOB-X, but in the case of AAL5 (which 381 we require -- see below) it can be used interchangeably and 382 consistently with the above two capabilities. 384 In TM/UNI 4.0 the service categories are: 386 Constant Bit Rate (CBR) 387 Real-time Variable Bit Rate (rtVBR) 388 Non-real-time Variable Bit Rate (nrtVBR) 389 Unspecified Bit Rate (UBR) 390 Available Bit Rate (ABR) 392 The first two of these are real-time services, so that rtVBR is new 393 to TM/UNI 4.0. The ABR service is also new to TM/UNI 4.0. UBR 394 exists in all specifications, except perhaps in name, through the 395 ``best effort'' indication flag and/or the QoS Class 0. 397 The encoding used in 4.0 is consistent with the earlier versions. 398 For example, the Service Category is indicated solely by the 399 combination of the Bearer Capability and the Best Effort indication 400 flag. 402 In principle, it is possible to support any foreseeable service 403 through the use of BCOB-A/CBR. This is because the CBR service is 404 equivalent to having a ``pipe'' with specified bandwidth/timing. 405 However, it may be desirable to make better use of the ATM network's 406 resources by using other, less demanding, services when available. 407 (See RFC 1821 for a discussion of this [15].) 409 2.1.1 Service Categories for Guaranteed Service 411 There are two possible mappings for GS: 413 CBR (BCOB-A) 414 rtVBR 416 GS requires real-time support, that is, timing is required. Thus in 417 UNI 3.x, the bearer class BCOB-A (or an equivalent BCOB-X 418 formulation) must be used. In TM/UNI 4.0 either CBR or rtVBR is 419 appropriate. In both cases, GS would use a value of CLR 420 appropriately low for the link (i.e., such that congestion losses are 421 dominated by losses due to bit errors). The use of rtVBR may 422 encourage recovery of allocated bandwidth left unused by a source. 423 It also accomm odates more bursty sources with a larger bucket 424 parameter, and permits the use of tagging for excess traffic (see 425 Section 2.2). 427 Neither the BCOB-C bearer class, nor nrtVBR, UBR, ABR are good 428 matches for the GS service. These provide no delay estimates and 429 cannot guarantee consistently low delay for every packet. 431 Specification of BCOB-A or CBR requires specification of a PCR. The 432 PCR should be specified as the the token bucket rate parameter, with 433 appropriate conversion from bytes to cells (accounting for overhead), 434 of the GS TSpec. For both of these, the network provides a nominal 435 clearing rate of PCR with jitter toleration (bucket size) CDVT, 436 specified in a network specific manner (see below). 438 Specification of rtVBR requires the specification of two rates, SCR 439 and PCR. This models bursty traffic with specified peak and average 440 rates. With rtVBR, it is appropriate to map the PCR to the line rate 441 of incoming traffic and the SCR to the GS TSpec bucket rate. The ATM 442 bucket sizes are CDVT, in a network specific manner, and CDVT+BT, 443 respectively for the PCR and SCR parameters (see below). 445 2.1.2 Service Categories for Controlled Load 447 There are three possible mappings for CLS: 449 CBR (BCOB-A) 450 ABR 451 nrtVBR (BCOB-C) 453 Note that under UNI 3.x, only the first and third choices are 454 applicable. The first, with a CBR/BCOB-A connection, provides a 455 higher level of QoS than is necessary, but it may be convenient to 456 simply allocate a fixed-rate ``pipe'', which should be ubiquitously 457 supported in ATM networks. However unless this is the only choice 458 available, this will probably be wasteful of network resources. 460 The ABR category with a positive MCR aligns with the CLS idea of 461 ``best effort with floor.'' The ATM network agrees to forward cells 462 with a rate of at least MCR, which should be directly converted from 463 the token bucket rate of the TSpec. The bucket size parameter 464 measures approximately the amount of buffer required at the IWF. 466 The nrtVBR/BCOB-C category can also be used. The rtVBR category can 467 be used, although the edge device must choose a value for CTD and CDV 468 as a matter of local policy. 470 The UBR category does not provide enough capability for Controlled 471 Load. The point of CLS is to allow an allocation of resources, which 472 is facilitated by the token bucket traffic descriptor, and is 473 unavailable in UBR. 475 2.1.3 Service Categories for Best Effort 477 All of the service categories have the capability to carry Best 478 Effort service, but the natural service category is UBR (or, in UNI 479 3.x, BCOB-C or BCOB-X, with the best effort indication set). A CBR 480 or rtVBR clearly could be used, and since the service is not real- 481 time, a nrtVBR connection could also be used. In these cases the 482 rate parameter used reflects a bandwidth allocation in support of the 483 edge device's best effort connectivity to the far edge router. It 484 would be normal for traffic from many source/destination pairs to be 485 aggregated on this connection; indeed, since Best Effort is the 486 default IP behavior, the individual flows are not necessarily 487 identified or accounted for. CBR may be a preferred solution in the 488 case where best effort traffic is sufficiently highly aggregated that 489 a simple fixed-rate pipe is efficient. Both CBR and nrt-VBR provide 490 bandwidth allocation which may be useful for billing purposes. An 491 ABR connection could similarly be used to support Best Effort 492 traffic. The support of data communications protocols such as TCP/IP 493 is the explicit purpose for which ABR was specifically designed. It 494 is conceivable that a separate ABR connection would be made for 495 different IP flows, although the normal case would probably have all 496 IP Best Effort traffic with a common egress router sharing a single 497 ABR connection. 499 The rt-VBR service category may be considered less suitable, simply 500 because both the real-time delay constraint and the use of SCR/BT add 501 unnecessary complexity. 503 See specifications from the IETF ion working group [10, 11] for 504 related work on support of Best Effort service with ATM. 506 2.2 Cell Loss Priority Bit, Tagging and Conformance Definitions 508 An ATM header carries the Cell Loss Priority (CLP) bit. Cells with 509 CLP=1 are said to be ``tagged'' and have lower priority. This 510 tagging may be done by the source, to indicate relative priority 511 within the VC, or by a switch, to indicate traffic in violation of 512 policing parameters. Options involving the use of tagging are 513 decided at call setup time. 515 A Conformance Definition is a rule that determines whether a cell is 516 conforming to the traffic descriptor of the VC. The conformance 517 definition is given in terms of a Generic Cell Rate Algorithm (GCRA), 518 also known as a "leaky bucket" algorithm, for CBR and VBR services. 519 (UBR and ABR have network-specific conformance definitions. Note, 520 the term "compliance" in ATM is used to describe the behavior of a 521 connection.) 523 The network may tag cells which are non-conforming, rather than 524 dropping them only if the VC is set up to request tagging and the 525 network supports the tagging option. When congestion occurs, a 526 switch must attempt to discard tagged cells in preference to the 527 discarding of CLP=0 cells. However, the mechanism for doing this is 528 completely implementation specific. Tagged cells are treated with a 529 behavior which is Best Effort in the sense that they are transported 530 when bandwidth is available, queued when buffers are available, and 531 dropped when the resources are overcommitted. 533 Since GS and CLS services require excess traffic to be treated as 534 Best Effort, the tagging option should always be chosen (if 535 supported) in the VC setup as a means of ``downgrading'' non- 536 conformant cells. However, the term ``best effort'' seems to be used 537 with two distinguishable meanings in the int-serv specs. The first 538 is that of a service class that, in some typical scheduler 539 implementations, would correspond to a separate queue. Placing 540 excess traffic in best effort in this sense would be giving it lower 541 delay priority. The other sense is more generic, meaning that the 542 network would make a best effort to transport the traffic. A 543 reasonable expectation is that a network with no contending traffic 544 would transport the packet, while a very congested network would drop 545 the packet. A packet that could be tagged with lower loss priority 546 (such as the ATM CLP bit) would be more likely to be dropped, but 547 would not normally be transported out of order with respect to the 548 conforming portion of the flow. Such a mechanism would agree with 549 the latter definition of best effort, but not the former. 551 In TM/UNI 4.0 tagging does not apply to the CBR or ABR services. 552 However, there are three conformance definitions of VBR service (for 553 both rtVBR and nrtVBR) to consider. In VBR, only the conformance 554 definition VBR.3 supports tagging and applies the GCRA with PCR to 555 the aggregate CLP=0+1 cells, and another GCRA with SCR to the CLP=0 556 cells. Thus this conformance definition should always be used in 557 support of IP integrated services. For UBR service, conformance 558 definition UBR.2 supports the use of tagging, but a CLP=1 cell does 559 not imply non-conformance; it may be a hint of network congestion. 561 Once an ATM connection is established, and the particular conformance 562 definition is determined, the resulting policing action is mandatory. 563 Since the conformance algorithm operates on cells, when mapping rates 564 and bucket sizes from IP services to corresponding ATM parameters, a 565 correction needs to be made (at call setup time) for the ATM 566 segmentation overhead. Unfortunately this overhead, as a ratio, 567 depends on packet length, with the overhead largest for small 568 packets. Thus the appropriate correction could be based on minimum 569 packet size, expected packet size, or otherwise in a network specific 570 manner, determined at the edge device IWF. See Section 4.1. 572 It is always better fo the IWF to tag cells when it can anticipate 573 that the ATM network would do so. This is because the IWF knows the 574 IP packet boundaries and can tag all of the cells corresponding to a 575 packet. If left to the ATM layer UPC, the network would inevitably 576 carry some cells of packets which are worthless, because some other 577 cells from those packet are dropped due to non-conformance. 578 Therefore, the IWF, knowing the VC GCRA parameters, should always 579 anticipate the cells which will be tagged by the ATM UPC and tag all 580 of the cells uniformly across each affected packet. 582 2.3 ATM Adaptation Layer 584 The AAL type 5 encoding must be used, as specified in RFC 1483 and 585 RFC 1755. AAL5 requires specification of the maximum SDU size in both 586 the forward and reverse directions. Both GS and CLS specify a maximum 587 packet size as part of the TSpec and this value shall be used as the 588 maximum SDU in each direction for unicast connections, but only in 589 one direction for point-to-multipoint connections, which are 590 unidirectional. When more than one flow aggregated into a single VC, 591 the TSpecs are merged to yield the largest packet size. In no case 592 can this exceed 65535 (or, of course, the MTU of the link). 594 2.4 Broadband Low Layer Information 596 The B-LLI Information Element is transferred transparently by the ATM 597 network between the edge devices and is used to specify the 598 encapsulation method. Multiple B-LLI IEs may be sent as part of 599 negotiation. The default encapsulation LLC/SNAP [16] must be 600 supported as specified in RFC 1577 and RFC 1755. Additional 601 encapsulations are discussed in RFC 1755 and we refer to the 602 discussion there. 604 2.5 Traffic Descriptors 606 The ATM traffic descriptor always contains specification of a peak 607 cell rate (PCR) (in each direction). For variable rate services it 608 also contains specification of a sustainable cell rate (SCR) and 609 maximum burst size (MBS). The SCR and MBS form a leaky bucket pair 610 (rate, depth), while the bucket depth parameter for PCR is CDVT. 611 Note that CDVT is not signaled explicitly, but is determined by the 612 network operator, and serves as a measure of the jitter imposed by 613 the network. 615 Since CDVT is not signaled, and is presumed to be small, the leaky 616 bucket traffic descriptor (TSpec) of the Internet service cannot 617 always be directly mapped into PCR/CDVT parameters. Additional 618 buffering is needed at the IWF to account for the depth of the 619 bucket. 621 The Burst Tolerance is related to MBS (see TM 4.0 for details). 622 Roughly, they are both expressions of the bucket depth parameter that 623 goes with SCR. The units of BT is time while the units of MBS is 624 cells. Since both SCR and MBS are signalled, they can be computed 625 directly from the IP layer traffic description. The specific manner 626 in which resources are allocated from the traffic description is 627 implementation specific. Note that when translating the traffic 628 parameters, the segmentation overhead and minimum policed unit need 629 to be taken into account (see Section 4.2 below). 631 In ATM UNI SIG 4.0 there are the notions of Alternative Traffic 632 Descriptors and Minimal Traffic Descriptors. Alternative Traffic 633 Descriptors enumerate other acceptable choices for traffic 634 descriptors and are not considered here. Minimal Traffic Descriptors 635 are used in ``negotiation,'' which refers to the specific way in 636 which an ATM connection is set up. Very roughly it works like this, 637 taking PCR as an example: A minimal PCR and a requested PCR are 638 signalled, the requested PCR being the usual item signalled, and the 639 minimal PCR being the absolute minimum that the source edge device 640 will accept. When sensing the existence of both minimal and 641 requested parameters, the intermediate switches along the path may 642 reduce the requested PCR to a ``comfortable'' level. This choice is 643 part of admission control, and is therefore implementation dependent. 644 If at any point the requested PCR falls below the minimal PCR then 645 the call is cleared. Minimal Traffic Descriptors can be used to 646 present an acceptable range for parameters and ensure a higher 647 likelihood of call admission. Whether anything more specific about 648 Minimal Traffic Descriptors needs to be said here is left for further 649 study (FFS). In general, our discussion of connection parameters 650 assumes the values resulting from successful connection setup. 652 The Best Effort indicator (used only with UBR) and Tagging indicators 653 are also part of the signaled information element (IE) containing the 654 traffic descriptor. In the UNI SIG 4.0 traffic descriptor IE there 655 is an additional parameter, the Frame Discard indicator (see Section 656 2.7). 658 2.5.1 Translating Traffic Descriptors for Guaranteed Service 660 For Guaranteed Service there is a peak rate, p, a source Tspec rate, 661 r_s, a receiver Tspec rate r_r, and an Rspec rate, R. The two Tspec 662 rates are intended to support receiver heterogeneity, in the sense 663 that different receivers can accept different rates representing 664 subsets of the sender's traffic. In this document we leave this 665 feature for further study (FFS), and assume the two Tspec rates are 666 always identical. The Tspec rate describes the traffic itself, and 667 is used for policing, while the Rspec rate (which cannot be smaller) 668 is the allocated service rate. A receiver increases R over r to 669 reduce the delay. 671 When mapping Guaranteed Service onto a rtVBR VC, the ATM traffic 672 descriptor parameters (PCR, SCR, MBS) can be set within the following 673 bounds: 675 R <= PCR <= min(p, line rate) 676 r <= SCR <= PCR 677 0 <= MBS <- b. 679 Note that a receiver can choose R > p to lower the delay. This 680 leaves the first equation somewhat subject to interpretation. If a 681 receiver chooses R > line rate, it seems clear that the admission 682 control would simply reject the reservation. 684 The edge device has a buffer preceding the ATM network which must be 685 sufficient to absorb bursts arriving faster than they can be admitted 686 into the ATM network. For example, parameters may be set as PCR = R, 687 SCR = r, MBS = b. The edge device buffer of size b would absorb a 688 burst sent at any IP-level peak rate. Although this buffer exists, 689 the ATM network must accept bursts at rate PCR, at least R, to ensure 690 that the edge device delay is no greater than b/R. Since this buffer 691 is not in the ATM network, its delay is not included in D_ATM. 693 For GS over CBR, the service rate is mapped to the PCR parameter, 694 using the same constraint for PCR given above. The edge device again 695 requires adequate buffering to accommodate the TSpec bucket depth and 696 ensure delay before entering the ATM network of no more than b/R. If 697 PCR is greater than R, the buffer requirement may be relaxed 698 accordingly. 700 2.5.2 Translating Traffic Descriptors for Controlled Load Service 702 Controlled Load service has a peak rate, p, a Tspec rate, r, and a 703 corresponding bucket depth parameter, b. The ATM traffic parameters 704 for nrtVBR service category are constrained by 706 r <= SCR <= PCR <= min(p, line rate) 707 0 <= MBS <- b. 709 For ABR VCs, the Tspec rate would be used to set the minimum cell 710 rate (MCR) parameter. The bucket depth parameter does not map 711 directly to a signalled ATM parameter, so the edge device must have a 712 buffer of at least b bytes. 714 For CBR, the Tspec rate sets a lower bound on PCR, and again, the 715 available buffering in the edge device must be adequate to 716 accommodate possible bursts. 718 2.5.3 Translating Traffic Descriptors for Best Effort Service 720 For Best Effort service, there is no traffic description. The UBR 721 service category allows negotiation of PCR, simply to allow the 722 source to discover the smallest physical bottleneck along the path. 724 2.6 QoS Classes and Parameters 726 In TM/UNI 4.0 the three QoS parameters may be individually signalled. 727 These parameters are the Cell Loss Ratio (CLR), Cell Transfer Delay 728 (CTD), and Cell Delay Variation (CDV). In UNI 3.x the setup message 729 includes only the QoS Class, which is essentially an index to a 730 network specific table of values for these three parameters. A 731 network provider may choose to associate other parameters, such as 732 Severely Errored Cell Block Ratio, but these are less well understood 733 and accepted compared to the basic loss, delay and jitter parameters 734 mentioned here. The ITU has recently included a standard set of 735 parameter values for a (small) number of QoS classes in the latest 736 version of Recommendation I.356, October 1996. The network provider 737 may choose to define further network-specific QoS classes in addition 738 to these. The problem of agreement between network providers as to 739 the definition of QoS classes is completely unaddressed to date. We 740 will adopt a convention expressed in UNI 3.x, that assumes that QoS 741 class 1 is appropriate for low-delay, low-loss CBR connections, and 742 QoS class 3 is appropriate for variable rate connections with loss 743 and delay roughly appropriate for non-real-time data applications. 744 Note that the QoS class definitions in the new I.356 version may not 745 align with this model. 747 Since no IP layer counterparts to these ATM QoS parameters exist in 748 any of the IP services, they must be set by policy of the edge 749 device. The QoS classes can be chosen relatively easily. QoS class 750 1 should be used with Guaranteed Service and QoS class 3 should be 751 used with Controlled Load Service. Best Effort Service always gets 752 QoS class 0, which is unspecified QoS by definition. There are two 753 issues which amount to the same thing: First, the choice of 754 individually signalled parameter values (under TM/UNI 4.0) for GS and 755 CLS is the edge device policy. The second issue is choosing 756 parameter values for the two QoS classes, which is the ATM network 757 policy. If the same network operator controls both, then these 758 problems are identical; if not, an agreement to make the values 759 identical would be extremely desirable. 761 Note that we have mapped QoS class 1 and 3 onto Guaranteed and 762 Controlled Load service respectively. This is regardless of what 763 service category is used. So when running CLS over a CBR pipe, it 764 would not be inappropriate to use QoS class 3. This leaves the delay 765 unspecified (or much looser than with QoS 1). These comments should 766 be taken as preliminary, as these issues are far from clear, and 767 industry consensus should be sought. 769 2.7 Additional Parameters -- Frame Discard Mode 771 In TM/UNI 4.0 ATM allows the user to choose a mode where a dropped 772 cell causes all cells up to the last remaining in the AAL5 PDU to be 773 also dropped. This improves efficiency and the behavior of end-to- 774 end protocols such as TCP, since the remaining cells of a damaged PDU 775 are useless to the receiver. For IP over ATM, Frame Discard should 776 always be used in both directions, if available, for all services. 778 3.0 Discussion of IP-IS Protocol Features 780 3.1 Handling of Excess Traffic 782 (Placeholder for text.) 784 Reiterate that whole packets should be tagged, See Section 785 2.2. 787 3.2 Use of AdSpec in Guaranteed Service with ATM 789 The AdSpec is a feature of Guaranteed Service which allows a receiver 790 to calculate the worst-case delay associated with a GS flow. Three 791 quantities, C, D, and MPL, are accumulated (by simple addition of 792 components, one for each network element) in the PATH message from 793 source to receiver. The resulting values can be different for each 794 unique receiver. The maximum delay is then found by 796 delay <= b/R + C/R + D + MPL 798 The Maximum Path Latency (MPL) includes propagation delay and any 799 other unavoidable system delays. (We neglect the effect of maximum 800 packet size and peak rate here; see the GS specification [8] for the 801 more detailed equation.) The service rate requested by the receiver, 802 R, can be greater than the sender's Tspec rate, r. The effect of the 803 larger R is to allocate more bandwidth and, through this equation, 804 lower the packet delay. The burst size, b, is the leaky bucket 805 parameter from the Tspec, and is not changed by the receiver in the 806 Rspec. 808 The values of C and D which a router advertise will depend on both 809 the particular packet scheduling algorithm used in the router, and 810 the characteristics of the subnet attached to the router. We assume 811 here that each router (or the source host) takes responsibility for 812 its downstream subnet only. If the subnet is a simple point-to-point 813 link, then the subnet-specific parts of C and D will account for the 814 link transmission rate and MTU. An ATM subnet is more complex. 816 The edge router will always have an internal packet scheduler, which 817 will contribute to C and D. For this discussion we consider only the 818 ATM subnet-specific components. We further assume that the ATM 819 network will be represented as a "pure delay" element, contributing a 820 component to D, but not to C. The reason for this is that C would 821 depend on details of the cell scheduling algorithm inside the ATM 822 switches, which is not known by the edge device, where the AdSpec 823 parameters are accumulated. (In the special case where the edge 824 device does have enough information to modify C, it would not be 825 precluded.) Generally the delay behavior of the whole ATM cloud may 826 be expressed abstractly as a fixed constant D_ATM. 828 Since the AdSpec values are incremented before any reservation is 829 made, the edge device must have some knowledge about the VC which 830 would be set up in case a reservation were made. This does not 831 really add to the complexity of the device, since it must also have 832 this information in order to make an intelligent VC setup request. 833 For example, the edge device may have a cached table with the 834 propagation delay and a reasonable additional delay budget, from 835 which it composes a value of CTD for the VC setup. The device may 836 learn such information through VC setup negotiation, and, indeed, 837 there may be no other way to obtain that information. However, it 838 seems reasonable that these values would be cached for later use when 839 new VCs to the same egress router need to be established. 841 Therefore, we will presume a table with values of MPL (which includes 842 propagation delay) and expected queueing delays for each possible 843 egress edge device. (How such a table is maintained is 844 implementation specific.) The latter quantity is simply D_ATM, the 845 value added to the AdSpec D term to account for the ATM network. 846 When a RESV message arrives, causing a VC to be set up, the requested 847 value for CTD should then be given by 849 CTD = D_ATM + MPL + S_ATM. 851 The last term, S_ATM is the portion of the slack term applied to the 852 ATM portion of the path. Recall that the slack term [8] is positive 853 when the receiver can afford more delay than that computed from the 854 AdSpec. The ATM edge device may take part (or all) of the slack term 855 to relax the delay constraint on the ATM VC. The distribution of 856 delay slack among the nodes and subnets is network specific. 858 An important detail to note is the relationship between the b/R term 859 of the (Internet) delay and the corresponding MBS/SCR in the ATM 860 network, when using a VBR VC. The term b/R accounts for the delay 861 experienced by the last byte of a burst, of size b, which encounters 862 a congested node. In the simple ideal case, where the scheduling 863 algorithm emulates a fixed rate server, at rate R, the delay of the 864 last byte is b/R. Once this occurs, the stream has been smoothed, 865 and such a delay will not occur at later congested nodes, as long as 866 they also serve at rate R. The form of the delay equation expresses 867 this ideal behavior with C and D acting as error terms. Now, since 868 the delay which smooths the burst can occur outside of the ATM cloud, 869 the b/R term cannot include any delay within the ATM cloud. However, 870 a burst of size MBS is permitted to enter the ATM network, and it may 871 be served at a rate no greater than SCR. We might reasonably expect 872 a queueing delay of MBS/SCR to occur at a congested ATM switch. If 873 the ATM network will impose this delay, then it must be included in 874 the value of D_ATM advertised. If the ATM network can increase its 875 bandwidth allocation (e.g., due to CTD being lower than MBS/SCR), to 876 decrease this delay, then this behavior should be reflected in the 877 value of D_ATM. So, the information from which the edge device 878 determines D_ATM must reflect an accurate abstraction of the actual 879 behavior of the ATM network. To the extent that D_ATM is approximate 880 (and it must be an upper bound on the actual delay), it reduces the 881 chance that the VC setup will succeed, and/or increases its cost. 883 4.0 Discussion of Miscellaneous Items 885 4.1 Units Conversion 887 In the integrated services domain, bucket sizes and rates are 888 measured in bytes and bytes/sec, respectively, whereas for ATM, they 889 are measured in cells and cells/sec. 891 Packets are segmented into 53 byte cells of which the first 5 bytes 892 are header information. For 894 B = number of Bytes, 895 C = number of cells, 897 a rough approximation between the token bucket parameters (rate and 898 bucket depth) is 900 C = B/48. 902 This is actually a lower bound on C and does not take into account 903 the extra padding at the end of a partially filled cell, or the 8 904 byte trailer in the last cell of an AAL5 encoding. The actual 905 relationship between the number of cells and bytes of one packet is 907 C = 1 + int(B/48) + x, 909 where x = 1 if B mod 48 > 41 910 0 otherwise. 912 where int() is the rounding down operation. The third term is 0 or 913 1 and is 1 only when the remainder of B/48 is 41 or more. (An 914 additional cell is needed because the 41 bytes plus 8 byte trailer 915 will not fit in a cell.) 916 The above formula is not particularly amenable to engineering 917 considerations. By equating the number of bytes before and after 918 segmentation we have 920 48 C = B + 8 + A, 922 where A is the additional padding used in the last 2 cells and has 923 the range 0 <= A <= 47. From this we obtain a number of useful 924 observations. 926 For example, if one believes that the packet lengths are uniformly 927 distributed mod 48, then on average, 48 C = B + 8 + 47/2, or C = B/48 928 + .65625. 930 We can also make use of the upper bound on A to state that 48 C <= B 931 + 55. This is true for any one packet. Considering the number of 932 bytes in a stream of P packets, we have 934 48 C <= B + 55 P. 936 The number of packets P may not be a readily available quantity. 937 However, in terms of the minimum policed unit m, we know that P * m 938 <= B. Hence P <= B/m and 48 C <= B ( 1 + 55/m). That is, 940 C <= B/48 * (1 + 55/m). 942 5.0 Summary of ATM VC Setup Parameters for Guaranteed Service 944 This section describes how to create ATM VCs appropriately matched 945 for Guaranteed Service. The key points differentiating among ATM 946 choices are that real-time timing is required, that the data flow may 947 have a variable rate, and that demotion of non-conforming traffic to 948 best effort is required to be in agreement with the definition of GS. 949 For this reason, we prefer an rtVBR service in which tagging is 950 supported. Another good match is to use CBR with special handling of 951 any non-conforming traffic. 953 Note, in all cases the encodings assume point to multipoint 954 connections, where the backward channel is not used. This is done to 955 be consistent with rsvp, which generally assumes a multicast 956 scenerio. If a specific situation does not involve multicast, then 957 the IWF may make use of the backward channel in a point-to-point VC, 958 provided that the QoS parameters are mapped consistently for the 959 service provided. 961 5.1 Encoding GS Using Real-Time VBR (ATM Forum TM/UNI 4.0) 963 AAL 964 Type 5 965 Forward CPCS-SDU Size parameter M of TSpec 966 Backward CPCS-SDU Size 0 967 SSCS Type 0 (Null SSCS) 969 Traffic Descriptor 970 Forward PCR CLP=0+1 Note 1 971 Backward PCR CLP=0+1 0 972 Forward SCR CLP=0 Note 1 973 Backward SCR CLP=0 0 974 Forward MBS (CLP=0) Note 1 975 Backward MBS (CLP=0) 0 976 BE indicator NOT included 977 Forward Frame Discard bit 1 978 Backward Frame Discard bit 1 979 Tagging Forward bit 1 (Tagging requested) 980 Tagging Backward bit 1 (Tagging requested) 982 Broadband Bearer Capability 983 Bearer Class 16 (BCOB-X) Note 2 984 ATM Transfer Capability 9 (Real time VBR) Note 3 985 Susceptible to Clipping 00 (bit encoding for Not 986 susceptible) 987 User Plane Configuration 01 (bit encoding for pt-to-mpt) 989 Broadband Low Layer Information 990 User Information Layer 2 991 Protocol 12 (ISO 8802/2) 992 User Information Layer 3 993 Protocol 11 (ISO/IEC TR 9577) Note 4 994 ISO/IEC TR 9577 IPI 204 996 QoS Class 997 QoS Class Forward 1 Note 5 998 QoS Class Backward 1 Note 5 1000 QoS Parameters Note 6 1001 Acceptable Forward CDV 1002 Acceptable Forward CLR 1003 Forward Max CTD 1005 Note 1: See discussion Section 2.5.1. 1006 Note 2: Value 3 (BCOB-C) can also be used. 1007 Note 3: The ATC value 19 is not used. The value 19 implies CLR 1008 objective applies to the aggregate CLP=0+1 stream and 1009 that does not give desirable treatment of excess 1010 traffic in the case of IP. 1011 Note 4: For QoS VCs supporting GS or CLS, the layer 3 protocol should 1012 be specified. For BE VCs, it can be left unspecified, allowing 1013 the VC to be shared by multiple protocols, following RFC 1755. 1014 Note 5: Cf ITU I.365 (Oct 1996) for new definition. 1015 Note 6: See section 2.6 for the values to be used The cumulative CDV 1016 is also provided, but it depends on local implementation, and 1017 not on values mapped from IP level service parameters. 1019 5.2 Encoding GS Using CBR (ATM Forum TM/UNI 4.0) 1021 It is also possible to support GS using a CBR ``pipe.'' The 1022 advantage of this is that CBR is probably supported; the disadvantage 1023 is that data flows may not fill the pipe (utilization loss) and there 1024 is no tagging option available. 1026 AAL 1027 Type 5 1028 Forward CPCS-SDU Size parameter M of TSpec 1029 Backward CPCS-SDU Size parameter M of TSpec 1030 SSCS Type 0 (Null SSCS) 1032 Traffic Descriptor 1033 Forward PCR 0 Note 1 1034 Backward PCR 0 1035 Forward PCR 0+1 Note 1 1036 Backward PCR 0+1 0 1037 BE indicator NOT included 1038 Forward Frame Discard bit 1 1039 Backward Frame Discard bit 1 1040 Tagging Forward bit 1 (Tagging requested) 1041 Tagging Backward bit 1 (Tagging requested) 1043 Broadband Bearer Capability 1044 Bearer Class 16 (BCOB-X) Note 2 1045 ATM Transfer Capability 5 (CBR) Note 3, 4 1046 Susceptible to Clipping 00 (bit encoding for Not 1047 susceptible) 1048 User Plane Configuration 01 (bit encoding for pt-to-mpt) 1050 Broadband Low Layer Information 1051 User Information Layer 2 1052 Protocol 12 (ISO 8802/2) 1053 User Information Layer 3 1054 Protocol 11 (ISO/IEC TR 9577) Note 5 1055 ISO/IEC TR 9577 IPI 204 1057 QoS Class 1058 QoS Class Forward 1 Note 6 1059 QoS Class Backward 1 Note 6 1061 QoS Parameters Note 7 1062 Acceptable Forward CDV 1063 Acceptable Forward CLR 1064 Forward Max CTD 1066 Note 1: See discussion Section 2.5.1. 1067 Note 2: Value 1 (BCOB-A) can also be used. 1068 Note 3: If bearer class A is chosen the ATC field must be absent. 1069 Note 4: The ATC value 7 is not used. The value 7 implies CLR 1070 objective applies to the aggregate CLP=0+1 stream and 1071 that does not give desirable treatment of excess 1072 traffic in the case of IP. 1073 Note 5: For QoS VCs supporting GS or CLS, the layer 3 protocol should 1074 be specified. For BE VCs, it can be left unspecified, allowing 1075 the VC to be shared by multiple protocols, following RFC 1755. 1076 Note 6: Cf ITU I.365 (Oct 1996) for new definition. 1077 Note 7: See section 2.6 for the values to be used The cumulative CDV 1078 is also provided, but it depends on local implementation, and 1079 not on values mapped from IP level service parameters. 1081 5.3 Encoding GS Using Non-Real-Time VBR (ATM Forum TM/UNI 4.0) 1083 The remaining ATM service categories, including nrtVBR, do not 1084 provide delay guarantees and cannot be recommended as the best fits. 1085 However in some circumstances, the best fits may not be available. 1087 If nrtVBR is used, no hard delay can be given. However by using a 1088 variable rate service with low utilization, delay may be 1089 `reasonable', but not controlled. The encoding of GS as nrtVBR is 1090 the same as that for CLS using nrtVBR, except that the Forward PCR 1091 would be derived from the Tspec peak rate. See Section 6.2 below. 1093 5.4 Encoding GS Using ABR (ATM Forum TM/UNI 4.0) 1095 This is a very unlikely combination. The objective of the ABR 1096 service is to provide `low' loss rates which, via flow control, can 1097 result in delays. The introduction of delays is contrary to the 1098 design objectives of GS. If ABR were used for GS, the VC parameters 1099 would follow as for CLS over ABR. See Section 6.1. 1101 5.5 Encoding GS Using UBR (ATM Forum TM/UNI 4.0) 1103 The UBR service is the default lowest common denominator of the 1104 services. It cannot provide delay or loss guarantees. However if it 1105 is used for GS, it will be encoded in the same way as Best Effort 1106 over UBR, with the exception that the PCR would be determined from 1107 the peak rate of the Tspec. See Section 5.1. 1109 5.6 Encoding GS Using ATM Forum UNI 3.0/3.1 Specifications 1111 It is not recommended to support GS using VBR for the following 1112 reasons. The Class C bearer class does not represent real-time 1113 behavior. Appendix F of UNI 3.1 specification precludes the 1114 specification of traffic type "VBR" with the timing requirement "End 1115 to End timing Required" in conjunction with bearer class X. 1117 It is possible to support GS using a CBR ``pipe.'' The following 1118 table specifies the support of GS using CBR. 1120 AAL 1121 Type 5 1122 Forward CPCS-SDU Size parameter M of TSpec 1123 Backward CPCS-SDU Size parameter M of TSpec 1124 Mode 1 (Message mode) Note 1 1125 SSCS Type 0 (Null SSCS) 1127 Traffic Descriptor 1128 Forward PCR 0 Note 2 1129 Backward PCR 0 1130 Forward PCR 0+1 Note 2 1131 Backward PCR 0+1 0 1132 BE indicator NOT included 1133 Tagging Forward bit 1 (Tagging requested) 1134 Tagging Backward bit 1 (Tagging requested) 1136 Broadband Bearer Capability 1137 Bearer Class 16 (BCOB-X) Note 3 1138 Traffic Type 001 (bit encoding for Constant Bit 1139 Rate) 1140 Timing Requirements 01 (bit encoding for Timing 1141 Required) 1142 Susceptible to Clipping 00 (bit encoding for Not 1143 susceptible) 1144 User Plane Configuration 01 (bit encoding for pt-to-mpt) 1146 Broadband Low Layer Information 1147 User Information Layer 2 1148 Protocol 12 (ISO 8802/2) 1149 User Information Layer 3 1150 Protocol 11 (ISO/IEC TR 9577) Note 4 1151 ISO/IEC TR 9577 IPI 204 1153 QoS Class 1154 QoS Class Forward 1 1155 QoS Class Backward 1 1157 QoS Parameters 1158 Parameters are implied by the QOS Class 1160 Note 1: Only included for UNI 3.0. 1161 Note 2: See discussion, Section 2.5.1. 1162 Note 3: Value 1 (BCOB-A) can also be used. If BCOB-A is used Traffic 1163 Type and Timing Requirements fields are not included. 1164 Note 4: For QoS VCs supporting GS or CLS, the layer 3 protocol should 1165 be specified. For BE VCs, it can be left unspecified, allowing 1166 the VC to be shared by multiple protocols, following RFC 1755. 1168 6.0 Summary of ATM VC Setup Parameters for Controlled Load Service 1170 This section describes how to create ATM VCs appropriately matched 1171 for Controlled Load. CLS traffic is partly delay tolerant and of 1172 variable rate. NrtVBR and ABR (for TM/UNI 4.0 only) are the possible 1173 choices in supporting CLS. 1175 Generally we prefer to use point-to-multipoint connections. However 1176 this is not yet available in ABR. Other than in ABR, the encodings 1177 assume a point-to-multipoint connection. For a unicast connection, 1178 the backward parameters would be equal to the forward parameters. 1180 6.1 Encoding CLS Using ABR (ATM Forum TM/UNI 4.0) 1182 AAL 1183 Type 5 1184 Forward CPCS-SDU Size parameter M of TSpec 1185 Backward CPCS-SDU Size parameter M of TSpec 1186 SSCS Type 0 (Null SSCS) 1188 Traffic Descriptor 1189 Forward PCR CLP=0+1 From line rate 1190 Backward PCR CLP=0+1 From line rate 1191 Forward MCR CLP 0+1 From TSpec token bucket rate 1192 Backward MCR CLP 0+1 From TSpec token bucket rate 1193 BE indicator NOT included 1194 Forward Frame Discard bit 1 1195 Backward Frame Discard bit 1 1196 Tagging Forward bit 0 (Tagging not requested) 1197 Tagging Backward bit 0 (Tagging not requested) 1199 Broadband Bearer Capability 1200 Bearer Class 16 (BCOB-X) Note 1 1201 ATM Transfer Capability 12 (ABR) 1202 Traffic Type 010 (Variable Bit Rate) 1203 Timing Requirements 10 (Timing Not Required) 1204 Susceptible to Clipping 00 (Not susceptible) 1205 User Plane Configuration 00 (For pt-to-pt) 1207 Broadband Low Layer Information 1208 User Information Layer 2 1209 Protocol 12 (ISO 8802/2) 1210 User Information Layer 3 1211 Protocol 11 (ISO/IEC TR 9577) Note 2 1212 ISO/IEC TR 9577 IPI 204 1214 QoS Class 1215 QoS Class Forward 3 Note 3 1216 QoS Class Backward 3 Note 3 1218 QoS Parameters Note 4 1219 Acceptable Forward CDV 1220 Acceptable Forward CLR 1221 Forward Max CTD 1223 ABR Setup Parameters Note 5 ABR 1224 Additional Parameters Note 5 1226 Note 1: Value 3 (BCOB-C) can also be used. 1227 Note 2: For QoS VCs supporting GS or CLS, the layer 3 protocol should 1228 be specified. For BE VCs, it can be left unspecified, allowing 1229 the VC to be shared by multiple protocols, following RFC 1755. 1230 Note 3: Cf ITU I.365 (Oct 1996) for new definition. 1231 Note 4: See section 2.6 for the values to be used. The cumulative CDV 1232 is also provided, but it depends on local implementation, and 1233 not on values mapped from IP level service parameters. 1234 Note 5: Discussion of these parameters is beyond the scope of this draft. 1236 6.2 Encoding CLS Using Non-Real-Time VBR (ATM Forum TM/UNI 4.0) 1238 AAL 1239 Type 5 1240 Forward CPCS-SDU Size parameter M of TSpec 1241 Backward CPCS-SDU Size 0 1242 SSCS Type 0 (Null SSCS) 1244 Traffic Descriptor 1245 Forward PCR CLP=0+1 From line rate 1246 Backward PCR CLP=0+1 0 1247 Forward SCR CLP=0 From TSpec token bucket rate 1248 Backward SCR CLP=0 0 1249 Forward MBS (CLP=0) From TSpec bucket size param 1250 Backward MBS (CLP=0) 0 1251 BE indicator NOT included 1252 Forward Frame Discard bit 1 1253 Backward Frame Discard bit 1 1254 Tagging Forward bit 1 (Tagging requested) 1255 Tagging Backward bit 1 (Tagging requested) 1257 Broadband Bearer Capability 1258 Bearer Class 16 (BCOB-X) Note 1 1259 ATM Transfer Capability 10 (Non-real time VBR) Note 2, 3 1260 Susceptible to Clipping 00 (bit encoding Not susceptible) 1261 User Plane Configuration 01 (bit encoding pt-to-mpt) 1263 Broadband Low Layer Information 1264 User Information Layer 2 1265 Protocol 12 (ISO 8802/2) 1266 User Information Layer 3 1267 Protocol 11 (ISO/IEC TR 9577) Note 4 1268 ISO/IEC TR 9577 IPI 204 1270 QoS Class 1271 QoS Class Forward 3 Note 5 1272 QoS Class Backward 3 Note 5 1274 QoS Parameters Note 6 1275 Acceptable Forward CDV 1276 Acceptable Forward CLR 1277 Forward Max CTD 1279 Note 1: Value 3 (BCOB-C) can also be used. 1280 Note 2: If bearer class C is used, the ATC field must be absent 1281 Note 3: The ATC value 11 is not used. The value 11 implies CLR 1282 objective applies to the aggregate CLP=0+1 stream and 1283 that does not give desirable treatment of excess 1284 traffic in the case of IP. 1285 Note 4: For QoS VCs supporting GS or CLS, the layer 3 protocol should 1286 be specified. For BE VCs, it can be left unspecified, allowing 1287 the VC to be shared by multiple protocols, following RFC 1755. 1288 Note 5: Cf ITU I.365 (Oct 1996) for new definition. 1289 Note 6: See section 2.6 for the values to be used. The cumulative CDV 1290 is also provided, but it depends on local implementation, and 1291 not on values mapped from IP level service parameters. 1293 6.3 Encoding CLS Using Real-Time VBR (ATM Forum TM/UNI 4.0) 1295 The encoding of CLS using rtVBR imposes a hard limit on the delay, 1296 which is specified as an end-to-end delay in the ATM network. This 1297 is more stringent than the CLS service specifies and may result in 1298 less utilization of the network. 1300 If rtVBR is used to encode CLS, then the encoding is essentially the 1301 same as that for GS. The exceptions are that the Forward PCR is 1302 derived from the line rate and probably a different value of the 1303 transit delay and CDV will be specified. See Section 3.1. 1305 6.4 Encoding CLS Using CBR (ATM Forum TM/UNI 4.0) 1307 The encoding of CLS using CBR is more stringent than using rtVBR 1308 since it does not take into account the variable rate of the data. 1309 Consequently there may be even lower utilization of the network. 1311 To use CBR for CLS, the same encoding as in Section 3.2 would be 1312 used. However a different set of values of the QoS parameters will 1313 likely be used. 1315 6.5 Encoding CLS Using UBR (ATM Forum TM/UNI 4.0) 1317 This encoding gives no QoS guarantees and would be done in the same 1318 way as for BE traffic. See Section 5.1. 1320 6.6 Encoding CLS Using Non-Real-Time VBR as in UNI 3.0/3.1 1321 Specifications 1323 AAL 1324 Type 5 1325 Forward CPCS-SDU Size parameter M of TSpec 1326 Backward CPCS-SDU Size 0 1327 Mode 1 (Message mode) Note 1 1328 SSCS Type 0 (Null SSCS) 1330 Traffic Descriptor 1331 Forward PCR CLP=0+1 From line rate 1332 Backward PCR CLP=0+1 0 1333 Forward SCR CLP=0 From TSpec token bucket rate 1334 Backward SCR CLP=0 0 1335 Forward MBS (CLP=0) From TSpec bucket size param 1336 Backward MBS (CLP=0) 0 1337 BE indicator NOT included 1338 Tagging Forward bit 1 (Tagging requested) 1339 Tagging Backward bit 1 (Tagging requested) 1341 Broadband Bearer Capability 1342 Bearer Class 16 (BCOB-X) Note 2 1343 Traffic Type 010 (bit encoding for Variable Bit 1344 Rate) 1345 Timing Requirements 00 (bit encoding for No Indication) 1346 Susceptible to Clipping 00 (bit encoding for Not 1347 susceptible) 1348 User Plane Configuration 01 (bit encoding for For pt-to-mpt) 1350 Broadband Low Layer Information 1351 User Information Layer 2 1352 Protocol 12 (ISO 8802/2) 1353 User Information Layer 3 1354 Protocol 11 (ISO/IEC TR 9577) Note 3 1355 ISO/IEC TR 9577 IPI 204 1357 QoS Class 1358 QoS Class Forward 3 1359 QoS Class Backward 3 1361 QoS Parameters 1362 Parameters are implied by the QOS Class 1364 Note 1: Only included for UNI 3.0. 1365 Note 2: Value 3 (BCOB-C) can also be used. If BCOB-C is used Traffic 1366 Type and Timing Requirements fields are not included. 1367 Note 3: For QoS VCs supporting GS or CLS, the layer 3 protocol should 1368 be specified. For BE VCs, it can be left unspecified, allowing 1369 the VC to be shared by multiple protocols, following RFC 1755. 1371 7.0 Summary of ATM VC Setup Parameters for Best Effort Service 1373 This section describes how to create ATM VCs appropriately matched for 1374 Best Effort. The BE service does not need a reservation of resources. 1376 The following subsections are for information only. See the IETF ION 1377 working group draft on ATM signalling support for IP over ATM using UNI 1378 4.0 [11] for recommendations. 1380 7.1 Encoding Best Effort Service Using UBR (ATM Forum TM/UNI 4.0) 1382 This section is for information only. For recommendation, see the 1383 IETF ION working group draft on ATM signalling support for IP over 1384 ATM using UNI 4.0 [11]. 1386 AAL 1387 Type 5 1388 Forward CPCS-SDU Size MTU of link 1389 Backward CPCS-SDU Size MTU of link 1390 SSCS Type 0 (Null SSCS) 1392 Traffic Descriptor 1393 Forward PCR CLP=0+1 From line rate 1394 Backward PCR CLP=0+1 0 1395 BE indicator included 1396 Forward Frame Discard bit 1 1397 Backward Frame Discard bit 1 1398 Tagging Forward bit 1 (Tagging requested) 1399 Tagging Backward bit 1 (Tagging requested) 1401 Broadband Bearer Capability 1402 Bearer Class 16 (BCOB-X) Note 1 1403 ATM Transfer Capability 10 (Non-real time VBR) Note 2 1404 Susceptible to Clipping 00 (bit encoding for Not susceptible) 1405 User Plane Configuration 01 (bit encoding for pt-to-mpt) 1407 Broadband Low Layer Information 1408 User Information Layer 2 1409 Protocol 12 (ISO 8802/2) 1410 User Information Layer 3 1411 Protocol 11 (ISO/IEC TR 9577) Note 3 1412 ISO/IEC TR 9577 IPI 204 1414 QoS Class 1415 QoS Class Forward 0 1416 QoS Class Backward 0 1418 Note 1: Value 3 (BCOB-C) can also be used. 1419 Note 2: If bearer class C is used, the ATC field must be absent 1420 Note 3: For QoS VCs supporting GS or CLS, the layer 3 protocol should 1421 be specified. For BE VCs, it can be left unspecified, allowing 1422 the VC to be shared by multiple protocols, following RFC 1755. 1423 .fi 1425 7.2 Encoding Best Effort Service Using Other ATM Service Categories 1427 See the IETF ION working group draft on ATM signalling support for IP 1428 over ATM using UNI 4.0 [11]. 1430 8.0 Security 1432 Some security issues are raised in the rsvp specification [2], which 1433 would apply here as well. There are no additional security 1434 considerations raised in this document. 1436 9.0 Acknowledgements 1438 The authors would like to thank the members of the ISSLL working 1439 group for their input. In particular, thanks to Jon Bennett of Fore 1440 Systems, Roch Guerin of IBM and Susan Thomson of Bellcore. 1442 Appendix 1 Abbreviations 1444 AAL ATM Adaptation Layer 1445 ABR Available Bit Rate 1446 ATM Asynchronous Transfer Mode 1447 B-LLI Broadband Low Layer Information 1448 BCOB Broadband Connection-Oriented Bearer Capability 1449 BCOB-{A,C,X} Bearer Class A, C, or X 1450 BE Best Effort 1451 BT Burst Tolerance 1452 CBR Constant Bit Rate 1453 CDV Cell Delay Variation 1454 CDVT Cell Delay Variation Tolerance 1455 CLP Cell Loss Priority (bit) 1456 CLR Cell Loss Ratio 1457 CLS Controlled Load Service 1458 CPCS Common Part Convergence Sublayer 1459 CTD Cell Transfer Delay 1460 EOM End of Message 1461 FFS For Further Study 1462 GCRA Generic Cell Rate Algorithm 1463 GS Guaranteed Service 1464 IE Information Element 1465 IETF Internet Engineering Task Force 1466 IP Internet Protocol 1467 IS Integrated Services 1468 ISSLL Integrated Services over Specific Link Layers 1469 ITU International Telecommunication Union 1470 IWF Interworking Function 1471 LIJ Leaf Initiated Join 1472 LLC Logical Link Control 1473 MBS Maximum Burst Size 1474 MCR Minimum Cell Rate 1475 MPL Minimum Path Latency 1476 MTU Maximum Transfer Unit 1477 nrtVBR Non-real-time VBR 1478 PCR Peak Cell Rate 1479 PDU Protocol Data Unit 1480 QoS Quality of Service 1481 RESV Reservation Message (of rsvp protocol) 1482 RFC Request for Comment 1483 RSVP Resource Reservation Protocol 1484 Rspec Reservation Specification 1485 rtVBR Real-time VBR 1486 SCR Sustained Cell Rate 1487 SDU Service Data Unit 1488 SIG ATM Signaling (ATM Forum document) 1489 SNAP Subnetwork Attachment Point 1490 SSCS Service-Specific Convergence Sub-layer 1491 Sw Switch 1492 TCP Transport Control Protocol 1493 TM Traffic Management 1494 TSpec Traffic Specification 1495 UBR Unspecified Bit Rate 1496 UNI User-Network Interface 1497 UPC Usage Parameter Control (ATM traffic policing function) 1498 VBR Variable Bit Rate 1499 VC (ATM) Virtual Connection 1501 REFERENCES 1503 [1] R. Braden, D. Clark and S. Shenker, "Integrated Services in the 1504 Internet Architecture: an Overview", RFC 1633, June 1994. 1506 [2] R. Braden, L. Zhang, S. Berson, S. Herzog and S. Jamin, 1507 "Resource ReSerVation Protocol (RSVP) - Version 1 Functional 1508 Specification", Internet Draft, May 1996, 1511 [3] The ATM Forum, "ATM User-Network Interface Specification, Ver- 1512 sion 3.0", Prentice Hall, Englewood Cliffs NJ, 1993. 1514 [4] The ATM Forum, "ATM User-Network Interface Specification, Ver- 1515 sion 3.1", Prentice Hall, Upper Saddle River NJ, 1995. 1517 [5] The ATM Forum, "ATM User-Network Interface (UNI) Signalling 1518 Specification, Version 4.0", Prentice Hall, Upper Saddle River 1519 NJ, specification finalized July 1996; expected publication, 1520 late 1996; available at ftp://ftp.atmforum.com/pub. 1522 [6] The ATM Forum, "ATM Traffic Management Specification, Version 1523 4.0", Prentice Hall, Upper Saddle River NJ; specification final- 1524 ized April 1996; expected publication, late 1996; available at 1525 ftp://ftp.atmforum.com/pub. 1527 [7] M. W. Garrett, "A Service Architecture for ATM: From Applica- 1528 tions to Scheduling", IEEE Network Mag., Vol. 10, No. 3, pp. 6- 1529 14, May 1996. 1531 [8] S. Shenker, C. Partridge and R. Guerin, "Specification of 1532 Guaranteed Quality of Service", Internet Draft, August 1996, 1533 1535 [9] J. Wroclawski, "Specification of the Controlled-Load Network 1536 Element Service", Internet Draft, August 1996, draft-ietf- 1537 intserv-ctrl-load-svc-03.txt 1539 [10] M. Perez, F. Liaw, A. Mankin, E. Hoffman, D. Grossman and A. 1540 Malis, "ATM Signaling Support for IP over ATM", RFC 1755, Febru- 1541 ary 1995. 1543 [11] M. Perez and A. Mankin, "ATM Signalling Support for IP over ATM 1544 - UNI 4.0 Update", Internet Draft, November 1996, 1547 [12] S. Berson, L. Berger, "IP Integrated Services with RSVP over 1548 ATM", Internet Draft, September 1996, 1551 [13] S. Shenker and J. Wroclawski, "Network Element Service Specifi- 1552 cation Template", Internet Draft, November 1995, 1555 [14] J. Wroclawski, "The Use of RSVP with IETF Integrated Services", 1556 Internet Draft, August 1996, 1558 [15] M. Borden, E. Crawley, B. Davie and S. Batsell, "Integration of 1559 Real-time Services in an IP-ATM Network Architecture", "IP 1560 Authentication Header", RFC 1821, August 1995. 1562 [16] J. Heinanen, "Multiprotocol Encapsulation over ATM Adaptation 1563 Layer 5", RFC 1483, July 1993. 1565 AUTHORS' ADDRESSES 1567 Mark W. Garrett Marty Borden 1568 Bellcore New Oak Communications, Inc. 1569 445 South Street 42 Nanog Park 1570 Morristown, NJ 07960 Acton MA, 01720 1571 USA USA 1573 phone: +1 201 829-4439 phone: +1 508 266-1011 1574 email: mwg@bellcore.com email: mborden@newoak.com 1575 Table of Contents