idnits 2.17.1 draft-thubert-6lo-forwarding-fragments-05.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- -- The draft header indicates that this document updates RFC4944, but the abstract doesn't seem to mention this, which it should. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year (Using the creation date from RFC4944, updated by this document, for RFC5378 checks: 2005-07-13) -- 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 (April 6, 2017) is 2576 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: 'RFC3031' is defined on line 607, but no explicit reference was found in the text -- Possible downref: Non-RFC (?) normative reference: ref. 'IEEE.802.15.4' == Outdated reference: A later version (-30) exists of draft-ietf-6tisch-architecture-11 -- Obsolete informational reference (is this intentional?): RFC 2309 (Obsoleted by RFC 7567) -- Obsolete informational reference (is this intentional?): RFC 5405 (Obsoleted by RFC 8085) Summary: 0 errors (**), 0 flaws (~~), 3 warnings (==), 6 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 6lo P. Thubert, Ed. 3 Internet-Draft Cisco Systems 4 Updates: 4944 (if approved) J. Hui 5 Intended status: Standards Track Nest Labs 6 Expires: October 8, 2017 April 6, 2017 8 LLN Fragment Forwarding and Recovery 9 draft-thubert-6lo-forwarding-fragments-05 11 Abstract 13 Considering that an LLN link-layer frame can have a payload below 100 14 bytes, an IPv6 packet might be fragmented more than 10 fragments at 15 the 6LoWPAN layer. In a 6LoWPAN mesh-under mesh network, the 16 fragments can be forwarded individually across the mesh, whereas a 17 route-over mesh network, a fragmented 6LoWPAN packet must be 18 reassembled at every hop, which causes latency and congestion. This 19 draft introduces a simple protocol to forward individual fragments 20 across a route-over mesh network, and, regardless of the type of 21 mesh, recover the loss of individual fragments across the mesh and 22 protect the network against bloat with a minimal flow control. 24 Status of This Memo 26 This Internet-Draft is submitted in full conformance with the 27 provisions of BCP 78 and BCP 79. 29 Internet-Drafts are working documents of the Internet Engineering 30 Task Force (IETF). Note that other groups may also distribute 31 working documents as Internet-Drafts. The list of current Internet- 32 Drafts is at http://datatracker.ietf.org/drafts/current/. 34 Internet-Drafts are draft documents valid for a maximum of six months 35 and may be updated, replaced, or obsoleted by other documents at any 36 time. It is inappropriate to use Internet-Drafts as reference 37 material or to cite them other than as "work in progress." 39 This Internet-Draft will expire on October 8, 2017. 41 Copyright Notice 43 Copyright (c) 2017 IETF Trust and the persons identified as the 44 document authors. All rights reserved. 46 This document is subject to BCP 78 and the IETF Trust's Legal 47 Provisions Relating to IETF Documents 48 (http://trustee.ietf.org/license-info) in effect on the date of 49 publication of this document. Please review these documents 50 carefully, as they describe your rights and restrictions with respect 51 to this document. Code Components extracted from this document must 52 include Simplified BSD License text as described in Section 4.e of 53 the Trust Legal Provisions and are provided without warranty as 54 described in the Simplified BSD License. 56 Table of Contents 58 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 2 59 2. Updating RFC 4944 . . . . . . . . . . . . . . . . . . . . . . 3 60 3. Terminology and Referenced Work . . . . . . . . . . . . . . . 4 61 4. New Dispatch types and headers . . . . . . . . . . . . . . . 5 62 4.1. Recoverable Fragment Dispatch type and Header . . . . . . 5 63 4.2. RFRAG Acknowledgment Dispatch type and Header . . . . . . 6 64 5. Fragments Recovery . . . . . . . . . . . . . . . . . . . . . 8 65 6. Forwarding Fragments . . . . . . . . . . . . . . . . . . . . 9 66 6.1. Upon the first fragment . . . . . . . . . . . . . . . . . 10 67 6.2. Upon the next fragments . . . . . . . . . . . . . . . . . 11 68 6.3. Upon the RFRAG Acknowledgments . . . . . . . . . . . . . 11 69 7. Security Considerations . . . . . . . . . . . . . . . . . . . 12 70 8. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 12 71 9. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 12 72 10. References . . . . . . . . . . . . . . . . . . . . . . . . . 12 73 10.1. Normative References . . . . . . . . . . . . . . . . . . 12 74 10.2. Informative References . . . . . . . . . . . . . . . . . 13 75 Appendix A. Rationale . . . . . . . . . . . . . . . . . . . . . 14 76 Appendix B. Requirements . . . . . . . . . . . . . . . . . . . . 16 77 Appendix C. Considerations On Flow Control . . . . . . . . . . . 17 78 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 18 80 1. Introduction 82 In most Low Power and Lossy Network (LLN) applications, the bulk of 83 the traffic consists of small chunks of data (in the order few bytes 84 to a few tens of bytes) at a time. Given that an IEEE Std. 802.15.4 85 [IEEE.802.15.4] frame can carry 74 bytes or more in all cases, 86 fragmentation is usually not required. However, and though this 87 happens only occasionally, a number of mission critical applications 88 do require the capability to transfer larger chunks of data, for 89 instance to support a firmware upgrades of the LLN nodes or an 90 extraction of logs from LLN nodes. In the former case, the large 91 chunk of data is transferred to the LLN node, whereas in the latter, 92 the large chunk flows away from the LLN node. In both cases, the 93 size can be on the order of 10K bytes or more and an end-to-end 94 reliable transport is required. 96 "Transmission of IPv6 Packets over IEEE 802.15.4 Networks" [RFC4944] 97 defines the original 6LoWPAN datagram fragmentation mechanism for 98 LLNs. One critical issue with this original design is that routing 99 an IPv6 packet across a route-over mesh requires to reassemble the 100 full packet at each hop, which may cause latency along a path and an 101 overall buffer bloat in the network. Those undesirable effects can 102 be alleviated by a hop-by-hop fragment forwarding technique such as 103 the one proposed in this specification, and arguably this could be 104 achieved without the need to define a new protocol. However, adding 105 that capability alone to the local implementation of the original 106 6LoWPAN fragmentation would not address the bulk of the issues raised 107 against it, and may create new issues like uncontrolled state in the 108 network. 110 Another issue against RFC 4944 [RFC4944] is that it does not define a 111 mechanism to first discover the loss of a fragment along a multi-hop 112 path (e.g. having exhausted the link-layer retries at some hop on the 113 way), and then to recover that loss. With RFC 4944, the forwarding 114 of a whole datagram fails when one fragment is not delivered properly 115 to the destination 6LoWPAN endpoint. End-to-end transport or 116 application-level mechanisms may require a full retransmission of the 117 datagram, wasting resources in an already constrained network. 119 In that situation, the source 6LoWPAN endpoint will not be aware that 120 a loss occurred and will continue sending all fragments for a 121 datagram that is already doomed. The original support is missing 122 signaling to abort a multi-fragment transmission at any time and from 123 either end, and, if the capability to forward fragments is 124 implemented, clean up the related state in the network. It is also 125 lacking flow control capabilities to avoid participating to a 126 congestion that may in turn cause the loss of a fragment and trigger 127 the retransmission of the full datagram. 129 This specification proposes a method to forward fragments across a 130 multi-hop route over mesh, and to recover individual fragments 131 between LLN endpoints. The method is designed to limit congestion 132 loss in the network and addresses the requirements that are detailed 133 in Appendix B. 135 2. Updating RFC 4944 137 This specification deprecates the fragmentation mechanism that is 138 specified in RFC 4944 [RFC4944] and replaces it with a model where 139 fragments can be forwarded end-to-end across a 6LoWPAN mesh network 140 of any type, and where fragments that are lost on the way can be 141 recovered individually. New dispatch types are defined in Section 4. 143 3. Terminology and Referenced Work 145 Past experience with fragmentation has shown that miss-associated or 146 lost fragments can lead to poor network behavior and, occasionally, 147 trouble at application layer. The reader is encouraged to read RFC 148 4963 [RFC4963] and follow the references for more information. 150 That experience led to the definition of "Path MTU discovery" 151 [RFC1191] (PMTUD) protocol that limits fragmentation over the 152 Internet. 154 Specifically in the case of UDP, valuable additional information can 155 be found in "UDP Usage Guidelines for Application Designers" 156 [RFC5405]. 158 Readers are expected to be familiar with all the terms and concepts 159 that are discussed in "IPv6 over Low-Power Wireless Personal Area 160 Networks (6LoWPANs): Overview, Assumptions, Problem Statement, and 161 Goals" [RFC4919] and "Transmission of IPv6 Packets over IEEE 802.15.4 162 Networks" [RFC4944]. 164 "The Benefits of Using Explicit Congestion Notification (ECN)" 165 [RFC8087] provides useful information on the potential benefits and 166 pitfalls of using ECN. 168 Quoting the "Multiprotocol Label Switching (MPLS) Architecture" 169 [RFC8087]: with MPLS, "packets are "labeled" before they are 170 forwarded. At subsequent hops, there is no further analysis of the 171 packet's network layer header. Rather, the label is used as an index 172 into a table which specifies the next hop, and a new label". That 173 technique is leveraged in this specification to forward fragments 174 that actually do not have a network layer header, since the 175 fragmentation occurs below IP. 177 This specification uses the following terms: 179 6LoWPAN endpoints 181 The LLN nodes in charge of generating or expanding a 6LoWPAN 182 header from/to a full IPv6 packet. The 6LoWPAN endpoints are the 183 points where fragmentation and reassembly take place. 185 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 186 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 187 document are to be interpreted as described in [RFC2119]. 189 4. New Dispatch types and headers 191 This specification aims at enabling to provide an MTU that is 192 equivalent to 2048 bytes to the upper layer, which can be the 6LoWPAN 193 Header Compression that is defined in the "Compression Format for 194 IPv6 Datagrams" [RFC6282] specification. In order to achieve this, 195 this specification enables the fragmentation and the reliable 196 transmission of fragments over a multihop 6LoWPAN mesh network. 198 This specification provides a technique that is derived from MPLS and 199 allows to forward fragments across a 6LoWPAN route-over mesh, but is 200 not needed in the mesh-under case. The datagram_tag is used as the 201 label and is locally unique to the node that is the MAC-layer source 202 of the fragment. There is thus no need for a global registry of 203 datagram_tags and a node may build the datagram_tag in its own 204 locally-significant way, as long as the resulting tag stays unique to 205 the particular datagram for the lifetime of that datagram. 207 This specification extends RFC 4944 [RFC4944] with 4 new Dispatch 208 types, for Recoverable Fragments (RFRAG) headers with or without 209 Acknowledgment Request (RFRAG vs. RFRAG-ARQ), and for the RFRAG 210 Acknowledgment back, with or without ECN Echo (RFRAG-ACK vs. RFRAG- 211 ECN). 213 (to be confirmed by IANA) The new 6LoWPAN Dispatch types use the 214 Value Bit Pattern of 11 1010xx, as follows: 216 Pattern Header Type 217 +------------+------------------------------------------+ 218 | 11 101000 | RFRAG - Recoverable Fragment | 219 | 11 101001 | RFRAG-ARQ - RFRAG with Ack Request | 220 | 11 101010 | RFRAG-ACK - RFRAG Acknowledgment | 221 | 11 101011 | RFRAG-ECN - RFRAG Ack with ECN Echo | 222 +------------+------------------------------------------+ 224 Figure 1: Additional Dispatch Value Bit Patterns 226 4.1. Recoverable Fragment Dispatch type and Header 228 In this specification, the size and offset of the fragments are 229 expressed on the compressed packet per as opposed to the uncompressed 230 - native packet - form. 232 The first fragment is recognized by a sequence of 0; it carries its 233 fragment_size and the datagram_size of the compressed packet, whereas 234 the other fragments carry their fragment_size and fragment_offset. 235 The last fragment for a datagram is recognized when its 236 fragment_offset and its fragment_size add up to the datagram_size. 238 Recoverable Fragments are sequenced and a bitmap is used in the RFRAG 239 Acknowledgment to indicate the received fragments by setting the 240 individual bits that correspond to their sequence. 242 1 2 3 243 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 1 244 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 245 |1 1 1 0 1 0 0 X|R|fragment_size| datagram_tag | 246 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 247 |sequence | fragment_offset | 248 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 249 X set == Ack Requested 251 Figure 2: RFRAG Dispatch type and Header 253 X: 1 bit; Ack Requested: when set, the sender requires an RFRAG 254 Acknowledgment from the receiver. 256 R: 1 bit; Reserved, MUST be set to 0 by the source and ignored by all 257 nodes. 259 Fragment_size: 7 bits unsigned integer. The size of this fragment 260 in units that depend on the MAC layer technology. For IEEE Std. 261 802.15.4, the unit is octet. 263 Sequence: 5 bits unsigned integer; the sequence number of the 264 fragment. Fragments are sequence numbered [0..N] where N is in 265 [0..31]. 267 Fragment_offset: 10 bits unsigned integer; when set to 0, this field 268 indicates an abort condition; else, its value depends on the value 269 of the Sequence. When the sequence is not 0, this field indicates 270 the offset of the fragment in the compressed form. When the 271 sequence is 0, denoting the first fragment of a datagram, this 272 field is overloaded to indicate the total_size of the compressed 273 packet, to help the receiver allocate an adapted buffer for the 274 reception and reassembly operations. 276 4.2. RFRAG Acknowledgment Dispatch type and Header 278 The specification also defines a 4-octet RFRAG Acknowledgment bitmap 279 that is used to confirm selectively the reception of individual 280 fragments. A given offset in the bitmap maps one to one with a given 281 sequence number. 283 The offset of the bit in the bitmap indicates which fragment is 284 acknowledged as follows: 286 1 2 3 287 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 1 288 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 289 | RFRAG Acknowledgment Bitmap | 290 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 291 ^ ^ 292 | | bitmap indicating whether: 293 | +--- Fragment with sequence 10 was received 294 +----------------------- Fragment with sequence 00 was received 296 Figure 3: RFRAG Acknowledgment bitmap encoding 298 Figure 4 shows an example Acknowledgment bitmap which indicates that 299 all fragments from sequence 0 to 20 were received, except for 300 fragments 1, 2 and 16 that were either lost or are still in the 301 network over a slower path. 303 1 2 3 304 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 1 305 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 306 |1|0|0|1|1|1|1|1|1|1|1|1|1|1|1|1|0|1|1|1|1|0|0|0|0|0|0|0|0|0|0|0| 307 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 309 Figure 4: Expanding 3 octets encoding 311 The RFRAG Acknowledgment Bitmap is included in a RFRAG Acknowledgment 312 header, as follows: 314 1 2 3 315 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 1 316 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 317 |1 1 1 0 1 0 1 Y| datagram_tag | 318 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 319 | RFRAG Acknowledgment Bitmap (32 bits) | 320 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 322 Figure 5: RFRAG Acknowledgment Dispatch type and Header 324 Y: 1 bit; Explicit Congestion Notification 326 When set, the sender indicates that at least one of the 327 acknowledged fragments was received with an Explicit Congestion 328 Notification, indicating that the path followed by the fragments 329 is subject to congestion. 331 RFRAG Acknowledgment Bitmap 332 An RFRAG Acknowledgment Bitmap, whereby but at offset x indicates 333 that fragment x was received. 335 5. Fragments Recovery 337 The Recoverable Fragment headers RFRAG and RFRAG-ARQ are used to 338 transport a fragment and optionally request an RFRAG Acknowledgment 339 that will confirm the good reception of a one or more fragments. An 340 RFRAG Acknowledgment can optionally carry an ECN indication; it is 341 carried as a standalone header in a message that is sent back to the 342 6LoWPAN endpoint that was the source of the fragments, as known by 343 its MAC address. The process ensures that at every hop, the source 344 MAC address and the datagram_tag in the received fragment are enough 345 information to send the RFRAG Acknowledgment back towards the source 346 6LoWPAN endpoint. 348 The 6LoWPAN endpoint that fragments the packets at 6LoWPAN level (the 349 sender) also controls the RFRAG Acknowledgments by setting the Ack 350 Requested flag in the RFRAG packets. It may set the Ack Requested 351 flag on any fragment so as to implement its own policy or perform 352 congestion control by limiting the number of fragments in the air, 353 IOW fragments that have been sent but for which reception or loss was 354 not positively confirmed by the other 6LoWPAN endpoint. When the 355 sender of the fragment knows that an underlying link-layer mechanism 356 protects the Fragments already it may refrain from using the RFRAG 357 Acknowledgment mechanism, and never set the Ack Requested bit. When 358 it receives a fragment with the ACK Request flag set, the 6LoWPAN 359 endpoint that reassembles the packets at 6LoWPAN level (the receiver) 360 sends back an RFRAG Acknowledgment to confirm reception of all the 361 fragments it has received so far, though it may slightly defer it to 362 let additional packets in. 364 The sender transfers a controlled number of fragments and MAY flag 365 the last fragment of a series with an RFRAG Acknowledgment Request. 366 The received MUST acknowledge a fragment with the acknowledgment 367 request bit set. If any fragment immediately preceding an 368 acknowledgment request is still missing, the receiver MAY 369 intentionally delay its acknowledgment to allow in-transit fragments 370 to arrive. delaying the acknowledgment might defeat the round trip 371 delay computation so it should be configurable and not enabled by 372 default. 374 The receiver interacts with the sender using an Acknowledgment 375 message with a bitmap that indicates which fragments were actually 376 received. The bitmap is a 32bit bitstring (a DWORD), which 377 accommodates up to 32 fragments and is sufficient to transport 2028 378 bytes over an IEEE Std. 802.15.4 MAC payload. For all n in [0..31], 379 bit n is set to 1 in the bitmap to indicate that fragment with 380 sequence n was received, otherwise the bit is set to 0. All 0s is a 381 NULL bitmap that indicates that the fragmentation process was 382 canceled by the receiver for that datagram. 384 The receiver MAY issue unsolicited acknowledgments. An unsolicited 385 acknowledgment enables the sender endpoint to resume sending if it 386 had reached its maximum number of outstanding fragments or indicate 387 that the receiver has cancelled the process of an individual 388 datagram. Note that acknowledgments might consume precious resources 389 so the use of unsolicited acknowledgments should be configurable and 390 not enabled by default. 392 The sender arms a retry timer to cover the fragment that carries the 393 Acknowledgment request. Upon time out, the sender assumes that all 394 the fragments on the way are received or lost. The process must have 395 completed within an acceptable time that is within the boundaries of 396 upper layer retries. The method detailed in [RFC6298] is recommended 397 for the computation of the retry timer. It is expected that the 398 upper layer retries obey the same or friendly rules in which case a 399 single round of fragment recovery should fit within the upper layer 400 recovery timers. 402 Fragments are sent in a round robin fashion: the sender sends all the 403 fragments for a first time before it retries any lost fragment; lost 404 fragments are retried in sequence, oldest first. This mechanism 405 enables the receiver to acknowledge fragments that were delayed in 406 the network before they are actually retried. 408 When the sender decides that a packet should be dropped and the 409 fragmentation process canceled, it sends a pseudo fragment with the 410 fragment_offset, sequence and fragment_size all set to 0, and no 411 data. Upon reception of this message, the receiver should clean up 412 all resources for the packet associated to the datagram_tag. If an 413 acknowledgment is requested, the receiver responds with a NULL 414 bitmap. 416 The receiver might need to cancel the process of a fragmented packet 417 for internal reasons, for instance if it is out of reassembly 418 buffers, or considers that this packet is already fully reassembled 419 and passed to the upper layer. In that case, the receiver SHOULD 420 indicate so to the sender with a NULL bitmap. Upon an acknowledgment 421 with a NULL bitmap, the sender MUST drop the datagram. 423 6. Forwarding Fragments 425 It is assumed that the first Fragment is large enough to carry the 426 IPv6 header and make routing decisions. If that is not so, then this 427 specification MUST NOT be used. 429 This specification enables intermediate routers to forward fragments 430 with no intermediate reconstruction of the entire packet. Upon the 431 first fragment, the routers lay an label along the path that is 432 followed by that fragment (that is IP routed), and all further 433 fragments are label switched along that path. As a consequence, 434 alternate routes not possible for individual fragments. The 435 datagram_tag is used to carry the label, that is swapped at each hop. 437 6.1. Upon the first fragment 439 In Route-Over mode, the MAC address changes at each hop. The label 440 that is formed and placed in the datagram_tag is associated to the 441 source MAC and only valid (and unique) for that source MAC. Say the 442 first fragment has: 444 Source IPv6 address = IP_A (maybe hops away) 446 Destination IPv6 address = IP_B (maybe hops away) 448 Source MAC = MAC_prv (prv as previous) 450 Datagram_tag= DT_prv 452 The intermediate router that forwards individual fragments does the 453 following: 455 a route lookup to get Next hop IPv6 towards IP_B, which resolves 456 as IP_nxt (nxt as next) 458 a MAC address resolution to get the MAC address associated to 459 IP_nxt, which resolves as MAC_nxt 461 Since it is a first fragment of a packet from that source MAC address 462 MAC_prv for that tag DT_prv, the router: 464 cleans up any leftover resource associated to the tupple (MAC_prv, 465 DT_prv) 467 allocates a new label for that flow, DT_nxt, from a Least Recently 468 Used pool or some similar procedure. 470 allocates a Label swap structure indexed by (MAC_prv, DT_prv) that 471 contains (MAC_nxt, DT_nxt) 473 allocates a Label swap structure indexed by (MAC_nxt, DT_nxt) that 474 contains (MAC_prv, DT_prv) 476 swaps the MAC info to from self to MAC_nxt 477 Swaps the datagram_tag to DT_nxt 479 At this point the router is all set and can forward the packet to 480 nxt. 482 6.2. Upon the next fragments 484 Upon next fragments (that are not first fragment), the router expects 485 to have already Label swap structure indexed by (MAC_prv, DT_prv). 486 The router: 488 lookups up the Label swap entry for (MAC_prv, DT_prv), which 489 resolves as (MAC_nxt, DT_nxt) 491 swaps the MAC info to from self to MAC_nxt; 493 Swaps the datagram_tag to DT_nxt 495 At this point the router is all set and can forward the packet to 496 nxt. 498 if the Label swap entry for (MAC_prv, DT_prv) is not found, the 499 router builds an RFRAG-ACK to indicate the error. The resulting 500 message has the following information: 502 MAC info set to from self to MAC_prv as found in the fragment 504 Swaps the datagram_tag set to DT_prv 506 Bitmap of all 0es to indicate the error 508 At this point the router is all set and can send the RFRAG-ACK back 509 ot the previous router. 511 6.3. Upon the RFRAG Acknowledgments 513 Upon an RFRAG Acknowledgment, the router expects to have already 514 Label swap structure indexed by (MAC_nxt, DT_nxt), which are 515 respectively the source MAC address of the received frame and the 516 received datagram_tag. DT_nxt should have been computed by this 517 router and this router should have assigned it to this particular 518 datagram. The router: 520 lookups up the Label swap entry for (MAC_nxt, DT_nxt), which 521 resolves as (MAC_prv, DT_prv) 523 swaps the MAC info to from self to MAC_prv; 524 Swaps the datagram_tag to DT_prv 526 At this point the router is all set and can forward the RFRAG-ACK to 527 prv. 529 if the Label swap entry for (MAC_nxt, DT_nxt) is not found, it simply 530 drops the packet. 532 if the RFRAG-ACK indicates either an error or that the fragment was 533 fully receive, the router schedules the Label swap entries for 534 recycling. If the RFRAG-ACK is lost on the way back, the source may 535 retry the last fragments, which will result as an error RFRAG-ACK 536 from the first router on the way that has already cleaned up. 538 7. Security Considerations 540 The process of recovering fragments does not appear to create any 541 opening for new threat compared to "Transmission of IPv6 Packets over 542 IEEE 802.15.4 Networks" [RFC4944]. 544 8. IANA Considerations 546 Need extensions for formats defined in "Transmission of IPv6 Packets 547 over IEEE 802.15.4 Networks" [RFC4944]. 549 9. Acknowledgments 551 The author wishes to thank Jay Werb, Christos Polyzois, Soumitri 552 Kolavennu, Pat Kinney, Margaret Wasserman, Richard Kelsey, Carsten 553 Bormann and Harry Courtice for their contributions and review. 555 10. References 557 10.1. Normative References 559 [IEEE.802.15.4] 560 IEEE, "IEEE Standard for Low-Rate Wireless Networks", 561 IEEE Standard 802.15.4, 562 . 564 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 565 Requirement Levels", BCP 14, RFC 2119, 566 DOI 10.17487/RFC2119, March 1997, 567 . 569 [RFC4944] Montenegro, G., Kushalnagar, N., Hui, J., and D. Culler, 570 "Transmission of IPv6 Packets over IEEE 802.15.4 571 Networks", RFC 4944, DOI 10.17487/RFC4944, September 2007, 572 . 574 [RFC6282] Hui, J., Ed. and P. Thubert, "Compression Format for IPv6 575 Datagrams over IEEE 802.15.4-Based Networks", RFC 6282, 576 DOI 10.17487/RFC6282, September 2011, 577 . 579 [RFC6298] Paxson, V., Allman, M., Chu, J., and M. Sargent, 580 "Computing TCP's Retransmission Timer", RFC 6298, 581 DOI 10.17487/RFC6298, June 2011, 582 . 584 10.2. Informative References 586 [I-D.ietf-6tisch-architecture] 587 Thubert, P., "An Architecture for IPv6 over the TSCH mode 588 of IEEE 802.15.4", draft-ietf-6tisch-architecture-11 (work 589 in progress), January 2017. 591 [RFC1191] Mogul, J. and S. Deering, "Path MTU discovery", RFC 1191, 592 DOI 10.17487/RFC1191, November 1990, 593 . 595 [RFC2309] Braden, B., Clark, D., Crowcroft, J., Davie, B., Deering, 596 S., Estrin, D., Floyd, S., Jacobson, V., Minshall, G., 597 Partridge, C., Peterson, L., Ramakrishnan, K., Shenker, 598 S., Wroclawski, J., and L. Zhang, "Recommendations on 599 Queue Management and Congestion Avoidance in the 600 Internet", RFC 2309, DOI 10.17487/RFC2309, April 1998, 601 . 603 [RFC2914] Floyd, S., "Congestion Control Principles", BCP 41, 604 RFC 2914, DOI 10.17487/RFC2914, September 2000, 605 . 607 [RFC3031] Rosen, E., Viswanathan, A., and R. Callon, "Multiprotocol 608 Label Switching Architecture", RFC 3031, 609 DOI 10.17487/RFC3031, January 2001, 610 . 612 [RFC3168] Ramakrishnan, K., Floyd, S., and D. Black, "The Addition 613 of Explicit Congestion Notification (ECN) to IP", 614 RFC 3168, DOI 10.17487/RFC3168, September 2001, 615 . 617 [RFC4919] Kushalnagar, N., Montenegro, G., and C. Schumacher, "IPv6 618 over Low-Power Wireless Personal Area Networks (6LoWPANs): 619 Overview, Assumptions, Problem Statement, and Goals", 620 RFC 4919, DOI 10.17487/RFC4919, August 2007, 621 . 623 [RFC4963] Heffner, J., Mathis, M., and B. Chandler, "IPv4 Reassembly 624 Errors at High Data Rates", RFC 4963, 625 DOI 10.17487/RFC4963, July 2007, 626 . 628 [RFC5405] Eggert, L. and G. Fairhurst, "Unicast UDP Usage Guidelines 629 for Application Designers", RFC 5405, 630 DOI 10.17487/RFC5405, November 2008, 631 . 633 [RFC5681] Allman, M., Paxson, V., and E. Blanton, "TCP Congestion 634 Control", RFC 5681, DOI 10.17487/RFC5681, September 2009, 635 . 637 [RFC7554] Watteyne, T., Ed., Palattella, M., and L. Grieco, "Using 638 IEEE 802.15.4e Time-Slotted Channel Hopping (TSCH) in the 639 Internet of Things (IoT): Problem Statement", RFC 7554, 640 DOI 10.17487/RFC7554, May 2015, 641 . 643 [RFC8087] Fairhurst, G. and M. Welzl, "The Benefits of Using 644 Explicit Congestion Notification (ECN)", RFC 8087, 645 DOI 10.17487/RFC8087, March 2017, 646 . 648 Appendix A. Rationale 650 There are a number of uses for large packets in Wireless Sensor 651 Networks. Such usages may not be the most typical or represent the 652 largest amount of traffic over the LLN; however, the associated 653 functionality can be critical enough to justify extra care for 654 ensuring effective transport of large packets across the LLN. 656 The list of those usages includes: 658 Towards the LLN node: 660 Packages of Commands: A number of commands or a full 661 configuration can by packaged as a single message to ensure 662 consistency and enable atomic execution or complete roll back. 663 Until such commands are fully received and interpreted, the 664 intended operation will not take effect. 666 Firmware update: For example, a new version of the LLN node 667 software is downloaded from a system manager over unicast or 668 multicast services. Such a reflashing operation typically 669 involves updating a large number of similar LLN nodes over a 670 relatively short period of time. 672 From the LLN node: 674 Waveform captures: A number of consecutive samples are measured 675 at a high rate for a short time and then transferred from a 676 sensor to a gateway or an edge server as a single large report. 678 Data logs: LLN nodes may generate large logs of sampled data for 679 later extraction. LLN nodes may also generate system logs to 680 assist in diagnosing problems on the node or network. 682 Large data packets: Rich data types might require more than one 683 fragment. 685 Uncontrolled firmware download or waveform upload can easily result 686 in a massive increase of the traffic and saturate the network. 688 When a fragment is lost in transmission, all fragments are resent, 689 further contributing to the congestion that caused the initial loss, 690 and potentially leading to congestion collapse. 692 This saturation may lead to excessive radio interference, or random 693 early discard (leaky bucket) in relaying nodes. Additional queuing 694 and memory congestion may result while waiting for a low power next 695 hop to emerge from its sleeping state. 697 Considering that [RFC4944] defines an MTU is 1280 bytes and that in 698 most incarnations (but 802.15.4G) a 802.15.4 frame can limit the MAC 699 payload to as few as 74 bytes, a packet might be fragmented into at 700 least 18 fragments at the 6LoWPAN shim layer. Taking into account 701 the worst-case header overhead for 6LoWPAN Fragmentation and Mesh 702 Addressing headers will increase the number of required fragments to 703 around 32. This level of fragmentation is much higher than that 704 traditionally experienced over the Internet with IPv4 fragments. At 705 the same time, the use of radios increases the probability of 706 transmission loss and Mesh-Under techniques compound that risk over 707 multiple hops. 709 Mechanisms such as TCP or application-layer segmentation could be 710 used to support end-to-end reliable transport. One option to support 711 bulk data transfer over a frame-size-constrained LLN is to set the 712 Maximum Segment Size to fit within the link maximum frame size. 713 Doing so, however, can add significant header overhead to each 714 802.15.4 frame. In addition, deploying such a mechanism requires 715 that the end-to-end transport is aware of the delivery properties of 716 the underlying LLN, which is a layer violation, and difficult to 717 achieve from the far end of the IPv6 network. 719 Appendix B. Requirements 721 For one-hop communications, a number of Low Power and Lossy Network 722 (LLN) link-layers propose a local acknowledgment mechanism that is 723 enough to detect and recover the loss of fragments. In a multihop 724 environment, an end-to-end fragment recovery mechanism might be a 725 good complement to a hop-by-hop MAC level recovery. This draft 726 introduces a simple protocol to recover individual fragments between 727 6LoWPAN endpoints that may be multiple hops away. The method 728 addresses the following requirements of a LLN: 730 Number of fragments 732 The recovery mechanism must support highly fragmented packets, 733 with a maximum of 32 fragments per packet. 735 Minimum acknowledgment overhead 737 Because the radio is half duplex, and because of silent time spent 738 in the various medium access mechanisms, an acknowledgment 739 consumes roughly as many resources as data fragment. 741 The new end-to-end fragment recovery mechanism should be able to 742 acknowledge multiple fragments in a single message and not require 743 an acknowledgment at all if fragments are already protected at a 744 lower layer. 746 Controlled latency 748 The recovery mechanism must succeed or give up within the time 749 boundary imposed by the recovery process of the Upper Layer 750 Protocols. 752 Support for out-of-order fragment delivery 754 Forwarding over a mesh network with rerouting and load balancing 755 can introduce out-of-sequence packets. 757 The recovery mechanism must account for packets that appear lost 758 but are actually only delayed over a different path. 760 Optional congestion control 761 The aggregation of multiple concurrent flows may lead to the 762 saturation of the radio network and congestion collapse. 764 The recovery mechanism should provide means for controlling the 765 number of fragments in transit over the LLN. 767 Appendix C. Considerations On Flow Control 769 Considering that a multi-hop LLN can be a very sensitive environment 770 due to the limited queuing capabilities of a large population of its 771 nodes, this draft recommends a simple and conservative approach to 772 congestion control, based on TCP congestion avoidance. 774 Congestion on the forward path is assumed in case of packet loss, and 775 packet loss is assumed upon time out. The draft allows to control 776 the number of outstanding fragments, that have been transmitted but 777 for which an acknowledgment was not received yet. It must be noted 778 that the number of outstanding fragments should not exceed the number 779 of hops in the network, but the way to figure the number of hops is 780 out of scope for this document. 782 Congestion on the forward path can also be indicated by an Explicit 783 Congestion Notification (ECN) mechanism. Though whether and how ECN 784 [RFC3168] is carried out over the LoWPAN is out of scope, this draft 785 provides a way for the destination endpoint to echo an ECN indication 786 back to the source endpoint in an acknowledgment message as 787 represented in Figure 5 in Section 4.2. 789 It must be noted that congestion and collision are different topics. 790 In particular, when a mesh operates on a same channel over multiple 791 hops, then the forwarding of a fragment over a certain hop may 792 collide with the forwarding of a next fragment that is following over 793 a previous hop but in a same interference domain. This draft enables 794 an end-to-end flow control, but leaves it to the sender stack to pace 795 individual fragments within a transmit window, so that a given 796 fragment is sent only when the previous fragment has had a chance to 797 progress beyond the interference domain of this hop. In the case of 798 6TiSCH [I-D.ietf-6tisch-architecture], which operates over the 799 TimeSlotted Channel Hopping [RFC7554] (TSCH) mode of operation of 800 IEEE802.14.5, a fragment is forwarded over a different channel at a 801 different time and it make full sense to fire a next fragment as soon 802 as the previous fragment has had its chance to be forwarded at the 803 next hop, retry (ARQ) operations included. 805 From the standpoint of a source 6LoWPAN endpoint, an outstanding 806 fragment is a fragment that was sent but for which no explicit 807 acknowledgment was received yet. This means that the fragment might 808 be on the way, received but not yet acknowledged, or the 809 acknowledgment might be on the way back. It is also possible that 810 either the fragment or the acknowledgment was lost on the way. 812 Because a meshed LLN might deliver frames out of order, it is 813 virtually impossible to differentiate these situations. In other 814 words, from the sender standpoint, all outstanding fragments might 815 still be in the network and contribute to its congestion. There is 816 an assumption, though, that after a certain amount of time, a frame 817 is either received or lost, so it is not causing congestion anymore. 818 This amount of time can be estimated based on the round trip delay 819 between the 6LoWPAN endpoints. The method detailed in [RFC6298] is 820 recommended for that computation. 822 The reader is encouraged to read through "Congestion Control 823 Principles" [RFC2914]. Additionally [RFC2309] and [RFC5681] provide 824 deeper information on why this mechanism is needed and how TCP 825 handles Congestion Control. Basically, the goal here is to manage 826 the amount of fragments present in the network; this is achieved by 827 to reducing the number of outstanding fragments over a congested path 828 by throttling the sources. 830 Section 5 describes how the sender decides how many fragments are 831 (re)sent before an acknowledgment is required, and how the sender 832 adapts that number to the network conditions. 834 Authors' Addresses 836 Pascal Thubert (editor) 837 Cisco Systems, Inc 838 Building D 839 45 Allee des Ormes - BP1200 840 MOUGINS - Sophia Antipolis 06254 841 FRANCE 843 Phone: +33 497 23 26 34 844 Email: pthubert@cisco.com 846 Jonathan W. Hui 847 Nest Labs 848 3400 Hillview Ave 849 Palo Alto, California 94304 850 USA 852 Email: jonhui@nestlabs.com