idnits 2.17.1 draft-ietf-lwig-tcp-constrained-node-networks-02.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 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.) ** There are 28 instances of too long lines in the document, the longest one being 11 characters in excess of 72. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year == The document doesn't use any RFC 2119 keywords, yet has text resembling RFC 2119 boilerplate text. -- The document date (February 27, 2018) is 2250 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- == Missing Reference: 'RFC 7228' is mentioned on line 668, but not defined == Unused Reference: 'I-D.ietf-core-coap-tcp-tls' is defined on line 892, but no explicit reference was found in the text ** Obsolete normative reference: RFC 793 (Obsoleted by RFC 9293) ** Obsolete normative reference: RFC 1323 (Obsoleted by RFC 7323) ** Obsolete normative reference: RFC 2460 (Obsoleted by RFC 8200) -- Obsolete informational reference (is this intentional?): RFC 1981 (Obsoleted by RFC 8201) -- Obsolete informational reference (is this intentional?): RFC 2385 (Obsoleted by RFC 5925) -- Obsolete informational reference (is this intentional?): RFC 2616 (Obsoleted by RFC 7230, RFC 7231, RFC 7232, RFC 7233, RFC 7234, RFC 7235) -- Obsolete informational reference (is this intentional?): RFC 7540 (Obsoleted by RFC 9113) Summary: 5 errors (**), 0 flaws (~~), 4 warnings (==), 5 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 LWIG Working Group C. Gomez 3 Internet-Draft UPC/i2CAT 4 Intended status: Informational J. Crowcroft 5 Expires: August 31, 2018 University of Cambridge 6 M. Scharf 7 Nokia 8 February 27, 2018 10 TCP Usage Guidance in the Internet of Things (IoT) 11 draft-ietf-lwig-tcp-constrained-node-networks-02 13 Abstract 15 This document provides guidance on how to implement and use the 16 Transmission Control Protocol (TCP) in Constrained-Node Networks 17 (CNNs), which are a characterstic of the Internet of Things (IoT). 18 Such environments require a lightweight TCP implementation and may 19 not make use of optional functionality. This document explains a 20 number of known and deployed techniques to simplify a TCP stack as 21 well as corresponding tradeoffs. The objective is to help embedded 22 developers with decisions on which TCP features to use. 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 https://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 August 31, 2018. 41 Copyright Notice 43 Copyright (c) 2018 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 (https://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 . . . . . . . . . . . . . . . . . . . . . . . . 3 59 2. Conventions used in this document . . . . . . . . . . . . . . 4 60 3. Characteristics of CNNs relevant for TCP . . . . . . . . . . 4 61 3.1. Network and link properties . . . . . . . . . . . . . . . 4 62 3.2. Usage scenarios . . . . . . . . . . . . . . . . . . . . . 5 63 3.3. Communication and traffic patterns . . . . . . . . . . . 6 64 4. TCP implementation and configuration in CNNs . . . . . . . . 6 65 4.1. Path properties . . . . . . . . . . . . . . . . . . . . . 6 66 4.1.1. Maximum Segment Size (MSS) . . . . . . . . . . . . . 7 67 4.1.2. Explicit Congestion Notification (ECN) . . . . . . . 7 68 4.1.3. Explicit loss notifications . . . . . . . . . . . . . 8 69 4.2. TCP guidance for small windows and buffers . . . . . . . 8 70 4.2.1. Single-MSS stacks - benefits and issues . . . . . . . 8 71 4.2.2. TCP options for single-MSS stacks . . . . . . . . . . 9 72 4.2.3. Delayed Acknowledgments for single-MSS stacks . . . . 9 73 4.2.4. RTO estimation for single-MSS stacks . . . . . . . . 10 74 4.3. General recommendations for TCP in CNNs . . . . . . . . . 10 75 4.3.1. Error recovery and congestion/flow control . . . . . 10 76 4.3.2. Selective Acknowledgments (SACK) . . . . . . . . . . 11 77 4.3.3. Delayed Acknowledgments . . . . . . . . . . . . . . . 11 78 5. TCP usage recommendations in CNNs . . . . . . . . . . . . . . 11 79 5.1. TCP connection initiation . . . . . . . . . . . . . . . . 12 80 5.2. TCP connection lifetime . . . . . . . . . . . . . . . . . 12 81 5.2.1. Long TCP connection lifetime . . . . . . . . . . . . 12 82 5.2.2. Short TCP connection lifetime . . . . . . . . . . . . 12 83 5.3. Number of parallel connections . . . . . . . . . . . . . 13 84 6. Security Considerations . . . . . . . . . . . . . . . . . . . 13 85 7. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 13 86 8. Annex. TCP implementations for constrained devices . . . . . 14 87 8.1. uIP . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 88 8.2. lwIP . . . . . . . . . . . . . . . . . . . . . . . . . . 14 89 8.3. RIOT . . . . . . . . . . . . . . . . . . . . . . . . . . 15 90 8.4. OpenWSN . . . . . . . . . . . . . . . . . . . . . . . . . 15 91 8.5. TinyOS . . . . . . . . . . . . . . . . . . . . . . . . . 15 92 8.6. FreeRTOS . . . . . . . . . . . . . . . . . . . . . . . . 16 93 8.7. uC/OS . . . . . . . . . . . . . . . . . . . . . . . . . . 16 94 8.8. Summary . . . . . . . . . . . . . . . . . . . . . . . . . 16 95 9. Annex. Changes compared to previous versions . . . . . . . . 18 96 9.1. Changes between -00 and -01 . . . . . . . . . . . . . . . 18 97 9.2. Changes between -01 and -02 . . . . . . . . . . . . . . . 18 98 10. References . . . . . . . . . . . . . . . . . . . . . . . . . 18 99 10.1. Normative References . . . . . . . . . . . . . . . . . . 18 100 10.2. Informative References . . . . . . . . . . . . . . . . . 20 101 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 23 103 1. Introduction 105 The Internet Protocol suite is being used for connecting Constrained- 106 Node Networks (CNNs) to the Internet, enabling the so-called Internet 107 of Things (IoT) [RFC7228]. In order to meet the requirements that 108 stem from CNNs, the IETF has produced a suite of new protocols 109 specifically designed for such environments (see e.g. 110 [I-D.ietf-lwig-energy-efficient]). New IETF protocol stack 111 components include the IPv6 over Low-power Wireless Personal Area 112 Networks (6LoWPAN) adaptation layer, the IPv6 Routing Protocol for 113 Low-power and lossy networks (RPL) routing protocol, and the 114 Constrained Application Protocol (CoAP). 116 As of the writing, the main current transport layer protocols in IP- 117 based IoT scenarios are UDP and TCP. However, TCP has been 118 criticized (often, unfairly) as a protocol for the IoT. In fact, 119 some TCP features are not optimal for IoT scenarios, such as 120 relatively long header size, unsuitability for multicast, and always- 121 confirmed data delivery. However, many typical claims on TCP 122 unsuitability for IoT (e.g. a high complexity, connection-oriented 123 approach incompatibility with radio duty-cycling, and spurious 124 congestion control activation in wireless links) are not valid, can 125 be solved, or are also found in well accepted IoT end-to-end 126 reliability mechanisms (see [IntComp] for a detailed analysis). 128 At the application layer, CoAP was developed over UDP [RFC7252]. 129 However, the integration of some CoAP deployments with existing 130 infrastructure is being challenged by middleboxes such as firewalls, 131 which may limit and even block UDP-based communications. This the 132 main reason why a CoAP over TCP specification has been developed 133 [RFC8323]. 135 Other application layer protocols not specifically designed for CNNs 136 are also being considered for the IoT space. Some examples include 137 HTTP/2 and even HTTP/1.1, both of which run over TCP by default 138 [RFC7540] [RFC2616], and the Extensible Messaging and Presence 139 Protocol (XMPP) [RFC6120]. TCP is also used by non-IETF application- 140 layer protocols in the IoT space such as the Message Queue Telemetry 141 Transport (MQTT) and its lightweight variants. 143 TCP is a sophisticated transport protocol that includes optional 144 functionality (e.g. TCP options) that may improve performance in 145 some environments. However, many optional TCP extensions require 146 complex logic inside the TCP stack and increase the codesize and the 147 RAM requirements. Many TCP extensions are not required for 148 interoperability with other standard-compliant TCP endpoints. Given 149 the limited resources on constrained devices, careful "tuning" of the 150 TCP implementation can make an implementation more lightweight. 152 This document provides guidance on how to implement and use TCP in 153 CNNs. The overarching goal is to offer simple measures to allow for 154 lightweight TCP implementation and suitable operation in such 155 environments. A TCP implementation following the guidance in this 156 document is intended to be compatible with a TCP endpoint that is 157 compliant to the TCP standards, albeit possibly with a lower 158 performance. This implies that such a TCP client would always be 159 able to connect with a standard-compliant TCP server, and a 160 corresponding TCP server would always be able to connect with a 161 standard-compliant TCP client. 163 This document assumes that the reader is familiar with TCP. A 164 comprehensive survey of the TCP standards can be found in [RFC7414]. 165 Similar guidance regarding the use of TCP in special environments has 166 been published before, e.g., for cellular wireless networks 167 [RFC3481]. 169 2. Conventions used in this document 171 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL","SHALL NOT", 172 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 173 document are to be interpreted as described in [RFC2119]. 175 3. Characteristics of CNNs relevant for TCP 177 3.1. Network and link properties 179 CNNs are defined in [RFC7228] as networks whose characteristics are 180 influenced by being composed of a significant portion of constrained 181 nodes. The latter are characterized by significant limitations on 182 processing, memory, and energy resources, among others [RFC7228]. 183 The first two dimensions pose constraints on the complexity and on 184 the memory footprint of the protocols that constrained nodes can 185 support. The latter requires techniques to save energy, such as 186 radio duty-cycling in wireless devices 187 [I-D.ietf-lwig-energy-efficient], as well as minimization of the 188 number of messages transmitted/received (and their size). 190 [RFC7228] lists typical network constraints in CNN, including low 191 achievable bitrate/throughput, high packet loss and high variability 192 of packet loss, highly asymmetric link characteristics, severe 193 penalties for using larger packets, limits on reachability over time, 194 etc. CNN may use wireless or wired technologies (e.g., Power Line 195 Communication), and the transmission rates are typically low (e.g. 196 below 1 Mbps). 198 For use of TCP, one challenge is that not all technologies in CNN may 199 be aligned with typical Internet subnetwork design principles 200 [RFC3819]. For instance, constrained nodes often use physical/link 201 layer technologies that have been characterized as 'lossy', i.e., 202 exhibit a relatively high bit error rate. Dealing with corruption 203 loss is one of the open issues in the Internet [RFC6077]. 205 3.2. Usage scenarios 207 There are different deployment and usage scenarios for CNNs. Some 208 CNNs follow the star topology, whereby one or several hosts are 209 linked to a central device that acts as a router connecting the CNN 210 to the Internet. CNNs may also follow the multihop topology 211 [RFC6606]. One key use case for the use of TCP is a model where 212 constrained devices connect to unconstrained servers in the Internet. 213 But it is also possible that both TCP endpoints run on constrained 214 devices. 216 In constrained environments, there can be different types of devices 217 [RFC7228]. For example, there can be devices with single combined 218 send/receive buffer, devices with a separate send and receive buffer, 219 or devices with a pool of multiple send/receive buffers. In the 220 latter case, it is possible that buffers also be shared for other 221 protocols. 223 When a CNN comprising one or more constrained devices and an 224 unconstrained device communicate over the Internet using TCP, the 225 communication possibly has to traverse a middlebox (e.g. a firewall, 226 NAT, etc.). Figure 1 illustrates such scenario. Note that the 227 scenario is asymmetric, as the unconstrained device will typically 228 not suffer the severe constraints of the constrained device. The 229 unconstrained device is expected to be mains-powered, to have high 230 amount of memory and processing power, and to be connected to a 231 resource-rich network. 233 Assuming that a majority of constrained devices will correspond to 234 sensor nodes, the amount of data traffic sent by constrained devices 235 (e.g. sensor node measurements) is expected to be higher than the 236 amount of data traffic in the opposite direction. Nevertheless, 237 constrained devices may receive requests (to which they may respond), 238 commands (for configuration purposes and for constrained devices 239 including actuators) and relatively infrequent firmware/software 240 updates. 242 +---------------+ 243 o o <-------- TCP communication -----> | | 244 o o | | 245 o o | Unconstrained | 246 o o +-----------+ | device | 247 o o o ------ | Middlebox | ------- | | 248 o o +-----------+ | (e.g. cloud) | 249 o o o | | 250 +---------------+ 251 constrained devices 253 Figure 1: TCP communication between a constrained device and an 254 unconstrained device, traversing a middlebox. 256 3.3. Communication and traffic patterns 258 IoT applications are characterized by a number of different 259 communication patterns. The following non-comprehensive list 260 explains some typical examples: 262 o Unidirectional transfers: An IoT device (e.g. a sensor) can send 263 (repeatedly) updates to the other endpoint. Not in every case 264 there is a need for an application response back to the IoT 265 device. 267 o Request-response patterns: An IoT device receiving a request from 268 the other endpoint, which triggers a response from the IoT device. 270 o Bulk data transfers: A typical example for a long file transfer 271 would be an IoT device firmware update. 273 A typical communication pattern is that a constrained device 274 communicates with an unconstrained device (cf. Figure 1). But it is 275 also possible that constrained devices communicate amongst 276 themselves. 278 4. TCP implementation and configuration in CNNs 280 This section explains how a TCP stack can deal with typical 281 constraints in CNN. The guidance in this section relates to the TCP 282 implementation and its configuration. 284 4.1. Path properties 285 4.1.1. Maximum Segment Size (MSS) 287 Some link layer technologies in the CNN space are characterized by a 288 short data unit payload size, e.g. up to a few tens or hundreds of 289 bytes. For example, the maximum frame size in IEEE 802.15.4 is 127 290 bytes. 6LoWPAN defined an adaptation layer to support IPv6 over IEEE 291 802.15.4 networks. The adaptation layer includes a fragmentation 292 mechanism, since IPv6 requires the layer below to support an MTU of 293 1280 bytes [RFC2460], while IEEE 802.15.4 lacked fragmentation 294 mechanisms. 6LoWPAN defines an IEEE 802.15.4 link MTU of 1280 bytes 295 [RFC4944]. Other technologies, such as Bluetooth LE [RFC7668], ITU-T 296 G.9959 [RFC7428] or DECT-ULE [RFC8105], also use 6LoWPAN-based 297 adaptation layers in order to enable IPv6 support. These 298 technologies do support link layer fragmentation. By exploiting this 299 functionality, the adaptation layers that enable IPv6 over such 300 technologies also define an MTU of 1280 bytes. 302 On the other hand, there exist technologies also used in the CNN 303 space, such as Master Slave / Token Passing (TP) [RFC8163], 304 Narrowband IoT (NB-IoT) [I-D.ietf-lpwan-overview] or IEEE 802.11ah 305 [I-D.delcarpio-6lo-wlanah], that do not suffer the same degree of 306 frame size limitations as the technologies mentioned above. The MTU 307 for MS/TP is recommended to be 1500 bytes [RFC8163], the MTU in NB- 308 IoT is 1600 bytes, and the maximum frame payload size for IEEE 309 802.11ah is 7991 bytes. 311 For the sake of lightweight implementation and operation, unless 312 applications require handling large data units (i.e. leading to an 313 IPv6 datagram size greater than 1280 bytes), it may be desirable to 314 limit the MTU to 1280 bytes in order to avoid the need to support 315 Path MTU Discovery [RFC1981]. 317 An IPv6 datagram size exceeding 1280 bytes can be avoided by setting 318 the TCP MSS not larger than 1220 bytes. (Note: IP version 6 is 319 assumed.) 321 4.1.2. Explicit Congestion Notification (ECN) 323 Explicit Congestion Notification (ECN) [RFC3168] may be used in CNNs. 324 ECN allows a router to signal in the IP header of a packet that 325 congestion is arising, for example when queue size reaches a certain 326 threshold. If such a packet encapsulates a TCP data packet, an ECN- 327 enabled TCP receiver will echo back the congestion signal to the TCP 328 sender by setting a flag in its next TCP ACK. The sender triggers 329 congestion control measures as if a packet loss had happened. In 330 that case, when the congestion window of a TCP sender has a size of 331 one segment, the TCP sender resets the retransmit timer, and will 332 only be able to send a new packet when the retransmit timer expires 334 [RFC3168]. Effectively, the TCP sender reduces at that moment its 335 sending rate from 1 segment per Round Trip Time (RTT) to 1 segment 336 per default RTO. 338 ECN can reduce packet losses, since congestion control measures can 339 be applied earlier than after the reception of three duplicate ACKs 340 (if the TCP sender window is large enough) or upon TCP sender RTO 341 expiration [RFC2884]. Therefore, the number of retries decreases, 342 which is particularly beneficial in CNNs, where energy and bandwidth 343 resources are typically limited. Furthermore, latency and jitter are 344 also reduced. 346 ECN is particularly appropriate in CNNs, since in these environments 347 transactional type interactions are a dominant traffic pattern. As 348 transactional data size decreases, the probability of detecting 349 congestion by the presence of three duplicate ACKs decreases. In 350 contrast, ECN can still activate congestion control measures without 351 requiring three duplicate ACKs. 353 4.1.3. Explicit loss notifications 355 There has been a significant body of research on solutions capable of 356 explicitly indicating whether a TCP segment loss is due to 357 corruption, in order to avoid activation of congestion control 358 mechanisms [ETEN] [RFC2757]. While such solutions may provide 359 significant improvement, they have not been widely deployed and 360 remain as experimental work. In fact, as of today, the IETF has not 361 standardized any such solution. 363 4.2. TCP guidance for small windows and buffers 365 This section discusses TCP stacks that focus on transferring a single 366 MSS. More general guidance is provided in Section 4.3. 368 4.2.1. Single-MSS stacks - benefits and issues 370 A TCP stack can reduce the RAM requirements by advertising a TCP 371 window size of one MSS, and also transmit at most one MSS of 372 unacknowledged data. In that case, both congestion and flow control 373 implementation is quite simple. Such a small receive and send window 374 may be sufficient for simple message exchanges in the CNN space. 375 However, only using a window of one MSS can significantly affect 376 performance. A stop-and-wait operation results in low throughput for 377 transfers that exceed the lengths of one MSS, e.g., a firmware 378 download. 380 If CoAP is used over TCP with the default setting for NSTART in 381 [RFC7252], a CoAP endpoint is not allowed to send a new message to a 382 destination until a response for the previous message sent to that 383 destination has been received. This is equivalent to an application- 384 layer window size of 1. For this use of CoAP, a maximum TCP window 385 of one MSS will be sufficient. 387 4.2.2. TCP options for single-MSS stacks 389 A TCP implementation needs to support options 0, 1 and 2 [RFC0793]. 390 These options are sufficient for interoperability with a standard- 391 compliant TCP endpoint, albeit many TCP stacks support additional 392 options and can negotiate their use. 394 A TCP implementation for a constrained device that uses a single-MSS 395 TCP receive or transmit window size may not benefit from supporting 396 the following TCP options: Window scale [RFC1323], TCP Timestamps 397 [RFC1323], Selective Acknowledgments (SACK) and SACK-Permitted 398 [RFC2018]. Also other TCP options may not be required on a 399 constrained device with a very lightweight implementation. 401 One potentially relevant TCP option in the context of CNNs is TCP 402 Fast Open (TFO) [RFC7413]. As described in Section 5.2.2, TFO can be 403 used to address the problem of traversing middleboxes that perform 404 early filter state record deletion. 406 4.2.3. Delayed Acknowledgments for single-MSS stacks 408 TCP Delayed Acknowledgments are meant to reduce the number of 409 transferred bytes within a TCP connection, but they may increase the 410 time until a sender may receive an ACK. There can be interactions 411 with stacks that use very small windows. 413 A device that advertises a single-MSS receive window should avoid use 414 of delayed ACKs in order to avoid contributing unnecessary delay (of 415 up to 500 ms) to the RTT [RFC5681], which limits the throughput and 416 can increase the data delivery time. 418 A device that can send at most one MSS of data is significantly 419 affected if the receiver uses delayed ACKs, e.g., if a TCP server or 420 receiver is outside the CNN. One known workaround is to split the 421 data to be sent into two segments of smaller size. A standard 422 compliant TCP receiver will then immediately acknowledge the second 423 segment, which can improve throughput. This "split hack" works if 424 the TCP receiver uses Delayed Acks, but the downside is the overhead 425 of sending two IP packets instead of one. 427 4.2.4. RTO estimation for single-MSS stacks 429 The Retransmission Timeout (RTO) estimation is one of the fundamental 430 TCP algorithms. There is a fundamental trade-off: A short, 431 aggressive RTO behavior reduces wait time before retransmissions, but 432 it also increases the probability of spurious timeouts. The latter 433 lead to unnecessary waste of potentially scarce resources in CNNs 434 such as energy and bandwidth. In contrast, a conservative timeout 435 can result in long error recovery times and thus needlessly delay 436 data delivery. 438 [RFC6298] describes the standard TCP RTO algorithm. If a TCP sender 439 uses very small window size and cannot use Fast Retransmit/Fast 440 Recovery or SACK, the Retransmission Timeout (RTO) algorithm has a 441 larger impact on performance than for a more powerful TCP stack. In 442 that case, RTO algorithm tuning may be considered, although careful 443 assessment of possible drawbacks is recommended. 445 As an example, an adaptive RTO algorithm for CoAP over UDP has been 446 defined [I-D.ietf-core-cocoa] that has been found to perform well in 447 CNN scenarios [Commag]. 449 4.3. General recommendations for TCP in CNNs 451 This section summarizes some widely used techniques to improve TCP, 452 with a focus on their use in CNNs. The TCP extensions discussed here 453 are useful in a wide range of network scenarios, including CNNs. 454 This section is not comprehensive. A comprehensive survey of TCP 455 extensions is published in [RFC7414]. 457 4.3.1. Error recovery and congestion/flow control 459 Devices that have enough memory to allow larger TCP window size can 460 leverage a more efficient error recovery using Fast Retransmit and 461 Fast Recovery [RFC5681]. These algorithms work efficiently for 462 window sizes of at least 5 MSS: If in a given TCP transmission of 463 segments 1,2,3,4,5, and 6 the segment 2 gets lost, the sender should 464 get an acknowledgement for segment 1 when 3 arrives and duplicate 465 acknowledgements when 4, 5, and 6 arrive. It will retransmit segment 466 2 when the third duplicate ack arrives. In order to have segment 2, 467 3, 4, 5, and 6 sent, the window has to be at least five. With an MSS 468 of 1220 byte, a buffer of the size of 5 MSS would require 6100 byte. 470 For bulk data transfers further TCP improvements may also be useful, 471 such as limited transmit [RFC3042]. 473 4.3.2. Selective Acknowledgments (SACK) 475 If a device with less severe memory and processing constraints can 476 afford advertising a TCP window size of several MSSs, it makes sense 477 to support the SACK option to improve performance. SACK allows a 478 data receiver to inform the data sender of non-contiguous data blocks 479 received, thus a sender (having previously sent the SACK-Permitted 480 option) can avoid performing unnecessary retransmissions, saving 481 energy and bandwidth, as well as reducing latency. SACK is 482 particularly useful for bulk data transfers. The receiver supporting 483 SACK will need to manage the reception of possible out-of-order 484 received segments, requiring sufficient buffer space. SACK adds 485 8*n+2 bytes to the TCP header, where n denotes the number of data 486 blocks received, up to 4 blocks. For a low number of out-of-order 487 segments, the header overhead penalty of SACK is compensated by 488 avoiding unnecessary retransmissions. 490 4.3.3. Delayed Acknowledgments 492 For certain traffic patterns, Delayed Acknowledgements may have a 493 detrimental effect, as already noted in Section 4.2.3. Advanced TCP 494 stacks may use heuristics to determine the maximum delay for an ACK. 495 For CNNs, the recommendation depends on the expected communication 496 patterns. 498 If a stack is able to deal with more than one MSS of data, it may 499 make sense to use a small timeout or disable delayed ACKs when 500 traffic over a CNN is expected to mostly be small messages with a 501 size typically below one MSS. For request-response traffic between a 502 constrained device and a peer (e.g. backend infrastructure) that uses 503 delayed ACKs, the maximum ACK rate of the peer will be typically of 504 one ACK every 200 ms (or even lower). If in such conditions the peer 505 device is administered by the same entity managing the constrained 506 device, it is recommended to disable delayed ACKs at the peer side. 508 In contrast, delayed ACKs allow to reduce the number of ACKs in bulk 509 transfer type of traffic, e.g. for firmware/software updates or for 510 transferring larger data units containing a batch of sensor readings. 512 5. TCP usage recommendations in CNNs 514 This section discusses how a TCP stack can be used by applications 515 that are developed for CNN scenarios. These remarks are by and large 516 independent of how TCP is exactly implemented. 518 5.1. TCP connection initiation 520 In the constrained device to unconstrained device scenario 521 illustrated above, a TCP connection is typically initiated by the 522 constrained device, in order for this device to support possible 523 sleep periods to save energy. 525 5.2. TCP connection lifetime 527 [[TODO: This section may need rewording in the next revision.]] 529 5.2.1. Long TCP connection lifetime 531 In CNNs, in order to minimize message overhead, a TCP connection 532 should be kept open as long as the two TCP endpoints have more data 533 to exchange or it is envisaged that further segment exchanges will 534 take place within an interval of two hours since the last segment has 535 been sent. A greater interval may be used in scenarios where 536 applications exchange data infrequently. 538 TCP keep-alive messages [RFC1122] may be supported by a server, to 539 check whether a TCP connection is active, in order to release state 540 of inactive connections. This may be useful for servers running on 541 memory-constrained devices. 543 Since the keep-alive timer may not be set to a value lower than two 544 hours [RFC1122], TCP keep-alive messages are not useful to guarantee 545 that filter state records in middleboxes such as firewalls will not 546 be deleted after an inactivity interval typically in the order of a 547 few minutes [RFC6092]. In scenarios where such middleboxes are 548 present, alternative measures to avoid early deletion of filter state 549 records (which might lead to frequent establishment of new TCP 550 connections between the two involved endpoints) include increasing 551 the initial value for the filter state inactivity timers (if 552 possible), and using application layer heartbeat messages. 554 5.2.2. Short TCP connection lifetime 556 A different approach to addressing the problem of traversing 557 middleboxes that perform early filter state record deletion relies on 558 using TFO [RFC7413]. In this case, instead of trying to maintain a 559 TCP connection for long time, possibly short-lived connections can be 560 opened between two endpoints while incurring low overhead. In fact, 561 TFO allows data to be carried in SYN (and SYN-ACK) packets, and to be 562 consumed immediately by the receceiving endpoint, thus reducing 563 overhead compared with the traditional three-way handshake required 564 to establish a TCP connection. 566 For security reasons, TFO requires the TCP endpoint that will open 567 the TCP connection (which in CNNs will typically be the constrained 568 device) to request a cookie from the other endpoint. The cookie, 569 with a size of 4 or 16 bytes, is then included in SYN packets of 570 subsequent connections. The cookie needs to be refreshed (and 571 obtained by the client) after a certain amount of time. 572 Nevertheless, TFO is more efficient than frequently opening new TCP 573 connections (by using the traditional three-way handshake) for 574 transmitting new data, as long as the cookie update rate is well 575 below the data new connection rate. 577 5.3. Number of parallel connections 579 [[TODO: This has been added in -02 but needs further alignment]] 581 TCP endpoints with a small amount of RAM may only support a small 582 number of connections. Each connection may result in overhead, and 583 depending on the internal TCP implementation, they may compete for 584 scarce resources. A careful application design may try to keep the 585 number of parallel connections as small as possible. 587 6. Security Considerations 589 Best current practise for securing TCP and TCP-based communication 590 also applies to CNN. As example, use of Transport Layer Security 591 (TLS) is strongly recommended if it is applicable. 593 There are also TCP options which can improve TCP security. Examples 594 include the TCP MD5 signature option [RFC2385] and the TCP 595 Authentication Option (TCP-AO) [RFC5925]. However, both options add 596 overhead and complexity. The TCP MD5 signature option adds 18 bytes 597 to every segment of a connection. TCP-AO typically has a size of 598 16-20 bytes. 600 For the mechanisms discussed in this document, the corresponding 601 considerations apply. For instance, if TFO is used, the security 602 considerations of [RFC7413] apply. 604 7. Acknowledgments 606 Carles Gomez has been funded in part by the Spanish Government 607 (Ministerio de Educacion, Cultura y Deporte) through the Jose 608 Castillejo grant CAS15/00336 and by European Regional Development 609 Fund (ERDF) and the Spanish Government through project 610 TEC2016-79988-P, AEI/FEDER, UE. Part of his contribution to this 611 work has been carried out during his stay as a visiting scholar at 612 the Computer Laboratory of the University of Cambridge. 614 The authors appreciate the feedback received for this document. The 615 following folks provided comments that helped improve the document: 616 Carsten Bormann, Zhen Cao, Wei Genyu, Ari Keranen, Abhijan 617 Bhattacharyya, Andres Arcia-Moret, Yoshifumi Nishida, Joe Touch, Fred 618 Baker, Nik Sultana, Kerry Lynn, Erik Nordmark, Markku Kojo, and 619 Hannes Tschofenig. Simon Brummer provided details on the RIOT TCP 620 implementation. Xavi Vilajosana provided details on the OpenWSN TCP 621 implementation. Rahul Jadhav provided details on the uIP TCP 622 implementation. 624 8. Annex. TCP implementations for constrained devices 626 This section overviews the main features of TCP implementations for 627 constrained devices. 629 8.1. uIP 631 uIP is a TCP/IP stack, targetted for 8 and 16-bit microcontrollers. 632 uIP has been deployed with Contiki and the Arduino Ethernet shield. 633 A code size of ~5 kB (which comprises checksumming, IP, ICMP and TCP) 634 has been reported for uIP [Dunk]. 636 uIP uses same buffer both incoming and outgoing traffic, with has a 637 size of a single packet. In case of a retransmission, an application 638 must be able to reproduce the same user data that had been 639 transmitted. 641 The MSS is announced via the MSS option on connection establishment 642 and the receive window size (of one MSS) is not modified during a 643 connection. Stop-and-wait operation is used for sending data. Among 644 other optimizations, this allows to avoid sliding window operations, 645 which use 32-bit arithmetic extensively and are expensive on 8-bit 646 CPUs. 648 Contiki uses the "split hack" technique (see Section 4.2.3) to avoid 649 delayed ACKs for senders using a single MSS. 651 8.2. lwIP 653 lwIP is a TCP/IP stack, targetted for 8- and 16-bit microcontrollers. 654 lwIP has a total code size of ~14 kB to ~22 kB (which comprises 655 memory management, checksumming, network interfaces, IP, ICMP and 656 TCP), and a TCP code size of ~9 kB to ~14 kB [Dunk]. 658 In contrast with uIP, lwIP decouples applications from the network 659 stack. lwIP supports a TCP transmission window greater than a single 660 segment, as well as buffering of incoming and outcoming data. Other 661 implemented mechanisms comprise slow start, congestion avoidance, 662 fast retransmit and fast recovery. SACK and Window Scale have been 663 recently added to lwIP. 665 8.3. RIOT 667 The RIOT TCP implementation (called GNRC TCP) has been designed for 668 Class 1 devices [RFC 7228]. The main target platforms are 8- and 669 16-bit microcontrollers. GNRC TCP offers a similar function set as 670 uIP, but it provides and maintains an independent receive buffer for 671 each connection. In contrast to uIP, retransmission is also handled 672 by GNRC TCP. GNRC TCP uses a single-MSS window size, which 673 simplifies the implementation. The application programmer does not 674 need to know anything about the TCP internals, therefore GNRC TCP can 675 be seen as a user-friendly uIP TCP implementation. 677 The MSS is set on connections establishment and cannot be changed 678 during connection lifetime. GNRC TCP allows multiple connections in 679 parallel, but each TCB must be allocated somewhere in the system. By 680 default there is only enough memory allocated for a single TCP 681 connection, but it can be increased at compile time if the user needs 682 multiple parallel connections. 684 The RIOT TCP implementation does not currently support classic POSIX 685 sockets. However, it supports an interface that has been inspired by 686 POSIX. 688 8.4. OpenWSN 690 The TCP implementation in OpenWSN is mostly equivalent to the uIP TCP 691 implementation. OpenWSN TCP implementation only supports the minimum 692 state machine functionality required. For example, it does not 693 perform retransmissions. 695 8.5. TinyOS 697 TinyOS has an experimental TCP stack that uses a simple nonblocking 698 library-based implementation of TCP, which provides a subset of the 699 socket interface primitives. The application is responsible for 700 buffering. The TCP library does not do any receive-side buffering. 701 Instead, it will immediately dispatch new, in-order data to the 702 application and otherwise drop the segment. A send buffer is 703 provided so that the TCP implementation can automatically retransmit 704 missing segments. Multiple TCP connections are possible. 706 8.6. FreeRTOS 708 FreeRTOS is a real-time operating system kernel for embedded devices 709 that is supported by 16- and 32-bit microprocessors. Its TCP 710 implementation is based on multiple-MSS window size, although a 711 'Tiny-TCP' option, which is a single-MSS variant, can be enabled. 712 Delayed ACKs are supported, with a 20-ms Delayed ACK timer as a 713 technique intended 'to gain performance'. 715 8.7. uC/OS 717 uC/OS is a real-time operating system kernel for embedded devices, 718 which is maintained by Micrium. uC/OS is intended for 8-, 16- and 719 32-bit microprocessors. The uC/OS TCP implementation supports a 720 multiple-MSS window size. 722 8.8. Summary 723 +---+---------+--------+----+-------+------+--------+-----+ 724 |uIP|lwIP orig|lwIP 2.0|RIOT|OpenWSN|TinyOS|FreeRTOS|uC/OS| 725 +------+-------------+---+---------+--------+----+-------+------+--------+-----+ 726 | |Data size(kB)| * | * | * | * | * | * | * | * | 727 |Memory+-------------+---+---------+--------+----+-------+------+--------+-----+ 728 | |Code size(kB)| <5|~9 to ~14| ~40 | * | * | * | <9.2 | * | 729 | | |(a)| (T1) | (b) | | | | (T2) | | 730 +------+-------------+---+---------+--------+----+-------+------+--------+-----+ 731 | |Win size(MSS)| 1 | Mult. | Mult. | 1 | 1 | Mult.| Mult. |Mult.| 732 | +-------------+---+---------+--------+----+-------+------+--------+-----+ 733 | | Slow start | No| Yes | Yes | No | No | Yes | * | Yes | 734 | T +-------------+---+---------+--------+----+-------+------+--------+-----+ 735 | C |Fast rec/retx| No| Yes | Yes | No | No | Yes | * | Yes | 736 | P +-------------+---+---------+--------+----+-------+------+--------+-----+ 737 | | Keep-alive | No| No | Yes | No | No | No | Yes | Yes | 738 | +-------------+---+---------+--------+----+-------+------+--------+-----+ 739 | f | Win. Scale | No| No | Yes | No | No | No | Yes | No | 740 | e +-------------+---+---------+--------+----+-------+------+--------+-----+ 741 | a | TCP timest. | No| No | Yes | No | No | No | Yes | No | 742 | t +-------------+---+---------+--------+----+-------+------+--------+-----+ 743 | u | SACK | No| No | Yes | No | No | No | Yes | No | 744 | r +-------------+---+---------+--------+----+-------+------+--------+-----+ 745 | e | Del. ACKs | No| Yes | Yes | No | No | No | Yes | Yes | 746 | s +-------------+---+---------+--------+----+-------+------+--------+-----+ 747 | | Socket | No| No |Optional|(I) | * |Subset| Yes | Yes | 748 | +-------------+---+---------+--------+----+-------+------+--------+-----+ 749 | |Concur. Conn.|Yes| Yes | Yes | Yes| Yes | Yes | * | * | 750 +------+-------------+---+---------+--------+----+-------+------+--------+-----+ 752 (T1) = TCP-only, on x86 and AVR platforms 753 (T2) = TCP-only, on ARM Cortex-M platform 754 (a) = includes IP, ICMP and TCP on x86 and AVR platforms 755 (b) = the whole protocol stack on mbed 756 (I) = interface inspired by POSIX 757 Mult. = Multiple 759 Figure 2: Summary of TCP features for differrent lightweight TCP 760 implementations. None of the implementations considered in this 761 Annex support ECN or TFO. 763 TODO: Add information about RAM requirements (in addition to 764 codesize) 766 9. Annex. Changes compared to previous versions 768 RFC Editor: To be removed prior to publication 770 9.1. Changes between -00 and -01 772 o Changed title and abstract 774 o Clarification that communcation with standard-compliant TCP 775 endpoints is required, based on feedback from Joe Touch 777 o Additional discussion on communication patters 779 o Numerous changes to address a comprehensive review from Hannes 780 Tschofenig 782 o Reworded security considerations 784 o Additional references and better distinction between normative and 785 informative entries 787 o Feedback from Rahul Jadhav on the uIP TCP implementation 789 o Basic data for the TinyOS TCP implementation added, based on 790 source code analysis 792 9.2. Changes between -01 and -02 794 o Added text to the Introduction section, and a reference, on 795 traditional bad perception of TCP for IoT 797 o Added sections on FreeRTOS and uC/OS 799 o Updated TinyOS section 801 o Updated summary table 803 o Reorganized Section 4 (single-MSS vs multiple-MSS window size), 804 some content now also in new Section 5 806 10. References 808 10.1. Normative References 810 [RFC0793] Postel, J., "Transmission Control Protocol", STD 7, 811 RFC 793, DOI 10.17487/RFC0793, September 1981, 812 . 814 [RFC1122] Braden, R., Ed., "Requirements for Internet Hosts - 815 Communication Layers", STD 3, RFC 1122, 816 DOI 10.17487/RFC1122, October 1989, 817 . 819 [RFC1323] Jacobson, V., Braden, R., and D. Borman, "TCP Extensions 820 for High Performance", RFC 1323, DOI 10.17487/RFC1323, May 821 1992, . 823 [RFC2018] Mathis, M., Mahdavi, J., Floyd, S., and A. Romanow, "TCP 824 Selective Acknowledgment Options", RFC 2018, 825 DOI 10.17487/RFC2018, October 1996, 826 . 828 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 829 Requirement Levels", BCP 14, RFC 2119, 830 DOI 10.17487/RFC2119, March 1997, 831 . 833 [RFC2460] Deering, S. and R. Hinden, "Internet Protocol, Version 6 834 (IPv6) Specification", RFC 2460, DOI 10.17487/RFC2460, 835 December 1998, . 837 [RFC3042] Allman, M., Balakrishnan, H., and S. Floyd, "Enhancing 838 TCP's Loss Recovery Using Limited Transmit", RFC 3042, 839 DOI 10.17487/RFC3042, January 2001, 840 . 842 [RFC3168] Ramakrishnan, K., Floyd, S., and D. Black, "The Addition 843 of Explicit Congestion Notification (ECN) to IP", 844 RFC 3168, DOI 10.17487/RFC3168, September 2001, 845 . 847 [RFC3819] Karn, P., Ed., Bormann, C., Fairhurst, G., Grossman, D., 848 Ludwig, R., Mahdavi, J., Montenegro, G., Touch, J., and L. 849 Wood, "Advice for Internet Subnetwork Designers", BCP 89, 850 RFC 3819, DOI 10.17487/RFC3819, July 2004, 851 . 853 [RFC5681] Allman, M., Paxson, V., and E. Blanton, "TCP Congestion 854 Control", RFC 5681, DOI 10.17487/RFC5681, September 2009, 855 . 857 [RFC5925] Touch, J., Mankin, A., and R. Bonica, "The TCP 858 Authentication Option", RFC 5925, DOI 10.17487/RFC5925, 859 June 2010, . 861 [RFC6298] Paxson, V., Allman, M., Chu, J., and M. Sargent, 862 "Computing TCP's Retransmission Timer", RFC 6298, 863 DOI 10.17487/RFC6298, June 2011, 864 . 866 [RFC7228] Bormann, C., Ersue, M., and A. Keranen, "Terminology for 867 Constrained-Node Networks", RFC 7228, 868 DOI 10.17487/RFC7228, May 2014, 869 . 871 [RFC7413] Cheng, Y., Chu, J., Radhakrishnan, S., and A. Jain, "TCP 872 Fast Open", RFC 7413, DOI 10.17487/RFC7413, December 2014, 873 . 875 10.2. Informative References 877 [Commag] A. Betzler, C. Gomez, I. Demirkol, J. Paradells, "CoAP 878 Congestion Control for the Internet of Things", IEEE 879 Communications Magazine, June 2016. 881 [Dunk] A. Dunkels, "Full TCP/IP for 8-Bit Architectures", 2003. 883 [ETEN] R. Krishnan et al, "Explicit transport error notification 884 (ETEN) for error-prone wireless and satellite networks", 885 Computer Networks 2004. 887 [I-D.delcarpio-6lo-wlanah] 888 Vega, L., Robles, I., and R. Morabito, "IPv6 over 889 802.11ah", draft-delcarpio-6lo-wlanah-01 (work in 890 progress), October 2015. 892 [I-D.ietf-core-coap-tcp-tls] 893 Bormann, C., Lemay, S., Tschofenig, H., Hartke, K., 894 Silverajan, B., and B. Raymor, "CoAP (Constrained 895 Application Protocol) over TCP, TLS, and WebSockets", 896 draft-ietf-core-coap-tcp-tls-11 (work in progress), 897 December 2017. 899 [I-D.ietf-core-cocoa] 900 Bormann, C., Betzler, A., Gomez, C., and I. Demirkol, 901 "CoAP Simple Congestion Control/Advanced", draft-ietf- 902 core-cocoa-03 (work in progress), February 2018. 904 [I-D.ietf-lpwan-overview] 905 Farrell, S., "LPWAN Overview", draft-ietf-lpwan- 906 overview-10 (work in progress), February 2018. 908 [I-D.ietf-lwig-energy-efficient] 909 Gomez, C., Kovatsch, M., Tian, H., and Z. Cao, "Energy- 910 Efficient Features of Internet of Things Protocols", 911 draft-ietf-lwig-energy-efficient-08 (work in progress), 912 October 2017. 914 [IntComp] C. Gomez, A. Arcia-Moret, J. Crowcroft, "TCP in the 915 Internet of Things: from ostracism to prominence", IEEE 916 Communications Magazine, January-February 2018. 918 [RFC1981] McCann, J., Deering, S., and J. Mogul, "Path MTU Discovery 919 for IP version 6", RFC 1981, DOI 10.17487/RFC1981, August 920 1996, . 922 [RFC2385] Heffernan, A., "Protection of BGP Sessions via the TCP MD5 923 Signature Option", RFC 2385, DOI 10.17487/RFC2385, August 924 1998, . 926 [RFC2616] Fielding, R., Gettys, J., Mogul, J., Frystyk, H., 927 Masinter, L., Leach, P., and T. Berners-Lee, "Hypertext 928 Transfer Protocol -- HTTP/1.1", RFC 2616, 929 DOI 10.17487/RFC2616, June 1999, 930 . 932 [RFC2757] Montenegro, G., Dawkins, S., Kojo, M., Magret, V., and N. 933 Vaidya, "Long Thin Networks", RFC 2757, 934 DOI 10.17487/RFC2757, January 2000, 935 . 937 [RFC2884] Hadi Salim, J. and U. Ahmed, "Performance Evaluation of 938 Explicit Congestion Notification (ECN) in IP Networks", 939 RFC 2884, DOI 10.17487/RFC2884, July 2000, 940 . 942 [RFC3481] Inamura, H., Ed., Montenegro, G., Ed., Ludwig, R., Gurtov, 943 A., and F. Khafizov, "TCP over Second (2.5G) and Third 944 (3G) Generation Wireless Networks", BCP 71, RFC 3481, 945 DOI 10.17487/RFC3481, February 2003, 946 . 948 [RFC4944] Montenegro, G., Kushalnagar, N., Hui, J., and D. Culler, 949 "Transmission of IPv6 Packets over IEEE 802.15.4 950 Networks", RFC 4944, DOI 10.17487/RFC4944, September 2007, 951 . 953 [RFC6077] Papadimitriou, D., Ed., Welzl, M., Scharf, M., and B. 954 Briscoe, "Open Research Issues in Internet Congestion 955 Control", RFC 6077, DOI 10.17487/RFC6077, February 2011, 956 . 958 [RFC6092] Woodyatt, J., Ed., "Recommended Simple Security 959 Capabilities in Customer Premises Equipment (CPE) for 960 Providing Residential IPv6 Internet Service", RFC 6092, 961 DOI 10.17487/RFC6092, January 2011, 962 . 964 [RFC6120] Saint-Andre, P., "Extensible Messaging and Presence 965 Protocol (XMPP): Core", RFC 6120, DOI 10.17487/RFC6120, 966 March 2011, . 968 [RFC6606] Kim, E., Kaspar, D., Gomez, C., and C. Bormann, "Problem 969 Statement and Requirements for IPv6 over Low-Power 970 Wireless Personal Area Network (6LoWPAN) Routing", 971 RFC 6606, DOI 10.17487/RFC6606, May 2012, 972 . 974 [RFC7252] Shelby, Z., Hartke, K., and C. Bormann, "The Constrained 975 Application Protocol (CoAP)", RFC 7252, 976 DOI 10.17487/RFC7252, June 2014, 977 . 979 [RFC7414] Duke, M., Braden, R., Eddy, W., Blanton, E., and A. 980 Zimmermann, "A Roadmap for Transmission Control Protocol 981 (TCP) Specification Documents", RFC 7414, 982 DOI 10.17487/RFC7414, February 2015, 983 . 985 [RFC7428] Brandt, A. and J. Buron, "Transmission of IPv6 Packets 986 over ITU-T G.9959 Networks", RFC 7428, 987 DOI 10.17487/RFC7428, February 2015, 988 . 990 [RFC7540] Belshe, M., Peon, R., and M. Thomson, Ed., "Hypertext 991 Transfer Protocol Version 2 (HTTP/2)", RFC 7540, 992 DOI 10.17487/RFC7540, May 2015, 993 . 995 [RFC7668] Nieminen, J., Savolainen, T., Isomaki, M., Patil, B., 996 Shelby, Z., and C. Gomez, "IPv6 over BLUETOOTH(R) Low 997 Energy", RFC 7668, DOI 10.17487/RFC7668, October 2015, 998 . 1000 [RFC8105] Mariager, P., Petersen, J., Ed., Shelby, Z., Van de Logt, 1001 M., and D. Barthel, "Transmission of IPv6 Packets over 1002 Digital Enhanced Cordless Telecommunications (DECT) Ultra 1003 Low Energy (ULE)", RFC 8105, DOI 10.17487/RFC8105, May 1004 2017, . 1006 [RFC8163] Lynn, K., Ed., Martocci, J., Neilson, C., and S. 1007 Donaldson, "Transmission of IPv6 over Master-Slave/Token- 1008 Passing (MS/TP) Networks", RFC 8163, DOI 10.17487/RFC8163, 1009 May 2017, . 1011 [RFC8323] Bormann, C., Lemay, S., Tschofenig, H., Hartke, K., 1012 Silverajan, B., and B. Raymor, Ed., "CoAP (Constrained 1013 Application Protocol) over TCP, TLS, and WebSockets", 1014 RFC 8323, DOI 10.17487/RFC8323, February 2018, 1015 . 1017 Authors' Addresses 1019 Carles Gomez 1020 UPC/i2CAT 1021 C/Esteve Terradas, 7 1022 Castelldefels 08860 1023 Spain 1025 Email: carlesgo@entel.upc.edu 1027 Jon Crowcroft 1028 University of Cambridge 1029 JJ Thomson Avenue 1030 Cambridge, CB3 0FD 1031 United Kingdom 1033 Email: jon.crowcroft@cl.cam.ac.uk 1035 Michael Scharf 1036 Nokia 1037 Lorenzstrasse 10 1038 Stuttgart, 70435 1039 Germany 1041 Email: michael.scharf@nokia.com