idnits 2.17.1 draft-bormann-coap-misc-21.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 a Security Considerations section. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year == Line 697 has weird spacing: '... code mid...' -- The document date (October 02, 2012) is 4224 days in the past. Is this intentional? -- Found something which looks like a code comment -- if you have code sections in the document, please surround them with '' and '' lines. Checking references for intended status: Informational ---------------------------------------------------------------------------- -- Looks like a reference, but probably isn't: '0' on line 1054 == Missing Reference: 'RFCXXXX' is mentioned on line 1415, but not defined == Unused Reference: 'CoRE201' is defined on line 323, but no explicit reference was found in the text == Outdated reference: A later version (-18) exists of draft-ietf-core-coap-12 == Outdated reference: A later version (-16) exists of draft-ietf-core-observe-06 == Outdated reference: A later version (-26) exists of draft-ietf-httpbis-p1-messaging-20 == Outdated reference: A later version (-26) exists of draft-ietf-httpbis-p4-conditional-20 == Outdated reference: A later version (-26) exists of draft-ietf-httpbis-p6-cache-20 ** Obsolete normative reference: RFC 2616 (Obsoleted by RFC 7230, RFC 7231, RFC 7232, RFC 7233, RFC 7234, RFC 7235) -- Obsolete informational reference (is this intentional?): RFC 5226 (Obsoleted by RFC 8126) Summary: 2 errors (**), 0 flaws (~~), 9 warnings (==), 4 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 CoRE Working Group C. Bormann 3 Internet-Draft K. Hartke 4 Intended status: Informational Universitaet Bremen TZI 5 Expires: April 5, 2013 October 02, 2012 7 Miscellaneous additions to CoAP 8 draft-bormann-coap-misc-21 10 Abstract 12 This short I-D makes a number of partially interrelated proposals how 13 to solve certain problems in the CoRE WG's main protocol, the 14 Constrained Application Protocol (CoAP). The current version has 15 been resubmitted to keep information about these proposals available; 16 the proposals are not all fleshed out at this point in time. 18 Status of this Memo 20 This Internet-Draft is submitted in full conformance with the 21 provisions of BCP 78 and BCP 79. 23 Internet-Drafts are working documents of the Internet Engineering 24 Task Force (IETF). Note that other groups may also distribute 25 working documents as Internet-Drafts. The list of current Internet- 26 Drafts is at http://datatracker.ietf.org/drafts/current/. 28 Internet-Drafts are draft documents valid for a maximum of six months 29 and may be updated, replaced, or obsoleted by other documents at any 30 time. It is inappropriate to use Internet-Drafts as reference 31 material or to cite them other than as "work in progress." 33 This Internet-Draft will expire on April 5, 2013. 35 Copyright Notice 37 Copyright (c) 2012 IETF Trust and the persons identified as the 38 document authors. All rights reserved. 40 This document is subject to BCP 78 and the IETF Trust's Legal 41 Provisions Relating to IETF Documents 42 (http://trustee.ietf.org/license-info) in effect on the date of 43 publication of this document. Please review these documents 44 carefully, as they describe your rights and restrictions with respect 45 to this document. Code Components extracted from this document must 46 include Simplified BSD License text as described in Section 4.e of 47 the Trust Legal Provisions and are provided without warranty as 48 described in the Simplified BSD License. 50 Table of Contents 52 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 3 53 2. Observing Resources in CoAP . . . . . . . . . . . . . . . . . 4 54 3. References . . . . . . . . . . . . . . . . . . . . . . . . . . 7 55 3.1. Normative References . . . . . . . . . . . . . . . . . . . 7 56 3.2. Informative References . . . . . . . . . . . . . . . . . . 8 57 Appendix A. The Nursery (Things that still need to ripen a 58 bit) . . . . . . . . . . . . . . . . . . . . . . . . 9 59 A.1. Envelope Options . . . . . . . . . . . . . . . . . . . . . 9 60 A.2. Payload-Length Option . . . . . . . . . . . . . . . . . . 10 61 A.3. URI Authorities with Binary Adresses . . . . . . . . . . . 10 62 A.4. Length-aware number encoding (o256) . . . . . . . . . . . 11 63 A.5. SMS encoding . . . . . . . . . . . . . . . . . . . . . . . 13 64 A.5.1. ASCII-optimized SMS encoding . . . . . . . . . . . . . 14 65 A.6. CONNECT . . . . . . . . . . . . . . . . . . . . . . . . . 16 66 A.6.1. Requesting a Tunnel with CONNECT . . . . . . . . . . . 17 67 A.6.2. Using a CONNECT Tunnel . . . . . . . . . . . . . . . . 17 68 A.6.3. Closing down a CONNECT Tunnel . . . . . . . . . . . . 18 69 Appendix B. The Museum (Things we did, but maybe not exactly 70 this way) . . . . . . . . . . . . . . . . . . . . . . 19 71 B.1. Getting rid of artificial limitations . . . . . . . . . . 19 72 B.1.1. Beyond 270 bytes in a single option . . . . . . . . . 19 73 B.1.2. Beyond 15 options . . . . . . . . . . . . . . . . . . 20 74 B.1.3. Implementing the option delimiter for 15 or more 75 options . . . . . . . . . . . . . . . . . . . . . . . 23 76 B.1.4. Option Length encoding beyond 270 bytes . . . . . . . 24 77 B.2. Registered Option . . . . . . . . . . . . . . . . . . . . 27 78 B.2.1. A Separate Suboption Number Space . . . . . . . . . . 27 79 B.2.2. Opening Up the Option Number Space . . . . . . . . . . 28 80 B.3. Enabling Protocol Evolution . . . . . . . . . . . . . . . 32 81 B.3.1. Potential new option number allocation . . . . . . . . 33 82 B.4. Patience, Leisure, and Pledge . . . . . . . . . . . . . . 35 83 B.4.1. Patience . . . . . . . . . . . . . . . . . . . . . . . 35 84 B.4.2. Leisure . . . . . . . . . . . . . . . . . . . . . . . 36 85 B.4.3. Pledge . . . . . . . . . . . . . . . . . . . . . . . . 36 86 B.4.4. Option Formats . . . . . . . . . . . . . . . . . . . . 37 87 Appendix C. The Cemetery (Things we won't do) . . . . . . . . . . 38 88 C.1. Example envelope option: solving #230 . . . . . . . . . . 38 89 C.2. Example envelope option: proxy-elective options . . . . . 39 90 C.3. Stateful URI compression . . . . . . . . . . . . . . . . . 39 91 Appendix D. Experimental Options . . . . . . . . . . . . . . . . 41 92 D.1. Options indicating absolute time . . . . . . . . . . . . . 41 93 D.2. Representing Durations . . . . . . . . . . . . . . . . . . 42 94 D.3. Rationale . . . . . . . . . . . . . . . . . . . . . . . . 43 95 D.4. Pseudo-Floating Point . . . . . . . . . . . . . . . . . . 44 96 D.5. A Duration Type for CoAP . . . . . . . . . . . . . . . . . 45 97 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . . 52 99 1. Introduction 101 The CoRE WG is tasked with standardizing an Application Protocol for 102 Constrained Networks/Nodes, CoAP [I-D.ietf-core-coap]. This protocol 103 is intended to provide RESTful [REST] services not unlike HTTP 104 [RFC2616], while reducing the complexity of implementation as well as 105 the size of packets exchanged in order to make these services useful 106 in a highly constrained network of themselves highly constrained 107 nodes. 109 This objective requires restraint in a number of sometimes 110 conflicting ways: 112 o reducing implementation complexity in order to minimize code size, 114 o reducing message sizes in order to minimize the number of 115 fragments needed for each message (in turn to maximize the 116 probability of delivery of the message), the amount of 117 transmission power needed and the loading of the limited-bandwidth 118 channel, 120 o reducing requirements on the environment such as stable storage, 121 good sources of randomness or user interaction capabilities. 123 This draft attempts to address a number of problems not yet 124 adequately solved in [I-D.ietf-core-coap]. The solutions proposed to 125 these problems are somewhat interrelated and are therefore presented 126 in one draft. As of the current version of the draft, the main body 127 is almost empty, since few significant problems remain with CoAP or 128 its satellite specifications. 130 The appendix contains the "CoAP cemetery" (Appendix C, possibly later 131 to move into its own draft), documenting roads that the WG decided 132 not to take, in order to spare readers from reinventing them in vain. 133 There is also a "CoAP museum" (Appendix B), which documents previous 134 forms of proposals part of which did make it into the main documents 135 in one form or another. Finally, the "CoAP nursery" (Appendix A) 136 contains half- to fully-baked proposals that might become interesting 137 as the basis for future extensions. 139 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 140 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 141 document are to be interpreted as described in [RFC2119]. 143 The term "byte" is used in its now customary sense as a synonym for 144 "octet". 146 2. Observing Resources in CoAP 148 (Co-Author for this section: Matthias Kovatsch) 150 There are two open issues related to -observe 151 [I-D.ietf-core-observe]: 153 o mixing freshness and observation lifetime, and 155 o non-cacheable resources. 157 To solve the first issue, we think that -observe should be clarified 158 as follows: 160 A server sends at least some notifications as confirmable messages. 161 Each confirmable notification is an opportunity for the server to 162 check if the client is still there. If the client acknowledges the 163 notification, it is assumed to be well and alive and still interested 164 in the resource. If it rejects the message with a reset message or 165 if it doesn't respond, it is assumed not longer to be interested and 166 is removed from the list of observers. So an observation 167 relationship can potentially go on forever, if the client 168 acknowledges each confirmable notification. If the server doesn't 169 send a notification for a while and wants to check if the client is 170 still there, it may send a confirmable notification with the current 171 resource state to check that. 173 So there is no mixing of freshness and lifetime going on. 175 The other issue is a bit less trivial to solve. The problem is that 176 normal CoAP and -observe actually have very different freshness 177 models: 179 Normally, when a client wants to know the current state of a 180 resource, it retrieves a representation, uses it and stores it in its 181 cache. Later, when it wants to know the current state again, it can 182 either use the stored representation provided that it's still fresh, 183 or retrieve a new representation, use it and store it in its cache. 185 If a server knows when the state of the resource will change the next 186 time, it can set the Max-Age of the representation to an accurate 187 time span. So the change of the resource state will coincide with 188 the expiration of the freshness of the representation stored in the 189 client's cache (ignoring network latency). 191 But if the resource changes its state unpredictably at any time, the 192 server can set the Max-Age only to an estimate. If the state then 193 actually changes before the freshness expires, the client wrongly 194 believes it has fresh information. Conversely, if the freshness 195 expires and the client wants to know the current state, the client 196 wrongly believes it has to make a new request although the 197 representation is actually still fresh (this is defused by ETag 198 validation). 200 -observe doesn't have these kinds of problems: the server does not 201 have to predict when the resource will change its state the next 202 time. It just sends a notification when it does. The new 203 representation invalidates the old representation stored in the 204 client's cache. So the client always has a fresh representation that 205 it can use when it wants to know the current resource state without 206 ever having to make a request. An explicit Max-Age is not needed for 207 determining freshness. 209 But -observe has a different set of problems: 211 The first problem is that the resource may change its state more 212 often than there is bandwidth available or the client can handle. 213 Thus, -observe cannot make any guarantee that a client will see every 214 state change. The solution is that -observe guarantees that the 215 client will eventually see the latest state change, and follows a 216 best effort approach to enable the client to see as many state 217 changes as possible. 219 The second problem is that, when a notification doesn't arrive for a 220 while, the client does not know if the resource did not change its 221 state or if the server lost its state and forgot that the client is 222 interested in the resource. We propose the following solution: With 223 each notification that the server sends, it makes a promise to send 224 another notification, and that it will send this next notification at 225 latest after a certain time span. This time span is included with 226 each notification. So when no notification arrives for a while and 227 the time span has not expired yet, the client assumes that the 228 resource did not change its state. If the time span has expired, no 229 notification has arrived and the client wants to know the current 230 state of the resource, it has to make a new request. 232 The third problem is that, when an intermediary is observing a 233 resource and wants to create a response from a representation stored 234 in its cache, it needs to specify a Max-Age. But the intermediary 235 cannot predict when it will receive the next notification, because 236 the next notification can arrive at any time. Unlike the origin 237 server, it also doesn't have the application-specific knowledge that 238 the origin server has. We propose the following solution: With each 239 notification a server sends, it includes a value that an intermediary 240 should use to calculate the Max-Age. 242 To summarize: 244 o A notification doesn't have a Max-Age; it's fresh until the next 245 notification arrives. A notification is the promise for another 246 notification that will arrive at latest after Next-Notification- 247 At-Latest. This value is included with every notification. The 248 promise includes that the server attempts to transmit a 249 notification to the client for the promised time span, even if the 250 client does not seem to respond, e.g., due to a temporary network 251 outage. 253 o A notification also contains another value, called Max-Age-Hint. 254 This value is used by a cache to calculate a Max-Age for the 255 representation if needed. In a cache, the Max-Age-Hint of a 256 representation is counted down like Max-Age. When it reaches 257 zero, however, the representation can be still used to satisfy 258 requests, but is non-cacheable (i.e., Max-Age is 0). The Max-Age- 259 Hint must be less than or equal to Next-Notification-At-Latest. 261 We see two possible ways to encode Next-Notification-At-Latest and 262 Max-Age-Hint in a message: 264 o The first way is to require the values of Next-Notification-At- 265 Latest and Max-Age-Hint to be the same, although they are 266 conceptually unrelated. Then, a single option in the message can 267 be used to hold both values. 269 o The second way is to include two options, one for Next- 270 Notification-At-Latest and one for Max-Age-Hint. Since Next- 271 Notification-At-Latest is less than or equal to Max-Age-Hint, the 272 first option should indicates Max-Age-Hint, and the second option 273 Next-Notification-At-Latest minus Max-Age-Hint with a default 274 value of 0. 276 3. References 278 3.1. Normative References 280 [I-D.ietf-core-coap] 281 Shelby, Z., Hartke, K., Bormann, C., and B. Frank, 282 "Constrained Application Protocol (CoAP)", 283 draft-ietf-core-coap-12 (work in progress), October 2012. 285 [I-D.ietf-core-observe] 286 Hartke, K., "Observing Resources in CoAP", 287 draft-ietf-core-observe-06 (work in progress), 288 September 2012. 290 [I-D.ietf-httpbis-p1-messaging] 291 Fielding, R., Lafon, Y., and J. Reschke, "HTTP/1.1, part 292 1: Message Routing and Syntax"", 293 draft-ietf-httpbis-p1-messaging-20 (work in progress), 294 July 2012. 296 [I-D.ietf-httpbis-p4-conditional] 297 Fielding, R., Lafon, Y., and J. Reschke, "HTTP/1.1, part 298 4: Conditional Requests", 299 draft-ietf-httpbis-p4-conditional-20 (work in progress), 300 July 2012. 302 [I-D.ietf-httpbis-p6-cache] 303 Fielding, R., Lafon, Y., Nottingham, M., and J. Reschke, 304 "HTTP/1.1, part 6: Caching", 305 draft-ietf-httpbis-p6-cache-20 (work in progress), 306 July 2012. 308 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 309 Requirement Levels", BCP 14, RFC 2119, March 1997. 311 [RFC2616] Fielding, R., Gettys, J., Mogul, J., Frystyk, H., 312 Masinter, L., Leach, P., and T. Berners-Lee, "Hypertext 313 Transfer Protocol -- HTTP/1.1", RFC 2616, June 1999. 315 [RFC4648] Josefsson, S., "The Base16, Base32, and Base64 Data 316 Encodings", RFC 4648, October 2006. 318 [RFC6256] Eddy, W. and E. Davies, "Using Self-Delimiting Numeric 319 Values in Protocols", RFC 6256, May 2011. 321 3.2. Informative References 323 [CoRE201] "Clarify use of retransmission window for duplicate 324 detection", CoRE ticket #201, 2012, 325 . 327 [CoRE214] "Adopt vendor-defined option into core-coap", CoRE 328 ticket #214, 2012, 329 . 331 [CoRE230] "Multiple Location options need to be processed as a 332 unit", CoRE ticket #230, 2012, 333 . 335 [CoRE241] "Proxy Safe & Cache Key indication for options", CoRE 336 ticket #241, 2012, 337 . 339 [REST] Fielding, R., "Architectural Styles and the Design of 340 Network-based Software Architectures", 2000. 342 [RFC1924] Elz, R., "A Compact Representation of IPv6 Addresses", 343 RFC 1924, April 1996. 345 [RFC2817] Khare, R. and S. Lawrence, "Upgrading to TLS Within 346 HTTP/1.1", RFC 2817, May 2000. 348 [RFC5198] Klensin, J. and M. Padlipsky, "Unicode Format for Network 349 Interchange", RFC 5198, March 2008. 351 [RFC5226] Narten, T. and H. Alvestrand, "Guidelines for Writing an 352 IANA Considerations Section in RFCs", BCP 26, RFC 5226, 353 May 2008. 355 [RFC6648] Saint-Andre, P., Crocker, D., and M. Nottingham, 356 "Deprecating the "X-" Prefix and Similar Constructs in 357 Application Protocols", BCP 178, RFC 6648, June 2012. 359 Appendix A. The Nursery (Things that still need to ripen a bit) 361 A.1. Envelope Options 363 As of [I-D.ietf-core-coap], options can take one of four types, two 364 of which are mostly identical: 366 o uint: A non-negative integer which is represented in network byte 367 order using a variable number of bytes (see [I-D.ietf-core-coap] 368 Appendix A); 370 o string: a sequence of bytes that is nominally a Net-Unicode string 371 [RFC5198]; 373 o opaque: a sequence of bytes. 375 o empty (not explicitly identified as a fourth type in 376 [I-D.ietf-core-coap]). 378 It turns out some options would benefit from some internal structure. 379 Also, it may be a good idea to be able to bundle multiple options 380 into one, in order to ensure consistency for a set of elective 381 options that need to be processed all or nothing (i.e., the option 382 becomes critical as soon as another option out of the set is 383 processed, too). 385 In this section, we introduce a fifth CoAP option type: Envelope 386 options. 388 An envelope option is a sequence of bytes that looks and is 389 interpreted exactly like a CoAP sequence of options. Instead of an 390 option count or an end-of-option marker, the sequence of options is 391 terminated by the end of the envelope option. 393 The nested options (options inside the envelope option) may come from 394 the same number space as the top-level CoAP options, or the envelope 395 option may define its own number space - this choice needs to be 396 defined for each envelope option. 398 If the top-level number space is used, the envelope option typically 399 will restrict the set of options that actually can be used in the 400 envelope. In particular, it is unlikely that an envelope option will 401 allow itself inside the envelope (this would be a recursive option). 403 Envelope options are a general, but simple mechanism. Some of its 404 potential uses are illustrated by two examples in the cemetery: 405 Appendix C.1 and Appendix C.2. (Each of these examples has its own 406 merits and demerits, which led us to decide not to pursue either of 407 them right now, but this should be discussed separately from the 408 concept of Envelope options employed in the examples.) 410 A.2. Payload-Length Option 412 Not all transport mappings may provide an unambiguous length of the 413 CoAP message. For UDP, it may also be desirable to pack more than 414 one CoAP message into one UDP payload (aggregation); in that case, 415 for all but the last message there needs to be a way to delimit the 416 payload of that message. 418 This can be solved using a new option, the Payload-Length option. If 419 this option is present, the value of this option is an unsigned 420 integer giving the length of the payload of the message (note that 421 this integer can be zero for a zero-length payload, which can in turn 422 be represented by a zero-length option value). (In the UDP 423 aggregation case, what would have been in the payload of this message 424 after "payload-length" bytes is then actually one or more additional 425 messages.) 427 A.3. URI Authorities with Binary Adresses 429 One problem with the way URI authorities are represented in the URI 430 syntax is that the authority part can be very bulky if it encodes an 431 IPv6 address in ASCII. 433 Proposal: Provide an option "Uri-Authority-Binary" that can be an 434 even number of bytes between 2 and 18 except 12 or 14. 436 o If the number of bytes is 2, the destination IP address of the 437 packet transporting the CoAP message is implied. 439 o If the number of bytes is 4 or 6, the first four bytes of the 440 option value are an IPv4 address in binary. 442 o If the number of bytes is 8 or 10, the first eight bytes are the 443 lower 64 bits of an IPv6 address; the upper eight bytes are 444 implied from the destination address of the packet transporting 445 the CoAP message. 447 o If the number of bytes is 16 or 18, the first 16 bytes are an IPv6 448 address. 450 o If two more bytes remain, this is a port number (as always in 451 network byte order). 453 The resulting authority is (conceptually translated into ASCII and) 454 used in place of an Uri-Authority option, or inserted into a Proxy- 455 Uri. Examples: 457 +-------------+------------------+---------+------------------------+ 458 | Proxy-Uri | Uri-Authority-Bi | Uri-Pat | URI | 459 | | nary | h | | 460 +-------------+------------------+---------+------------------------+ 461 | (none) | (none) | (none) | "/" | 462 | | | | | 463 | (none) | (none) | 'temp' | "/temp" | 464 | | | | | 465 | (none) | 2 bytes: 61616 | 'temp' | "coap://[DA]:61616/tem | 466 | | | | p" | 467 | | | | | 468 | (none) | 16 bytes: | temp | "coap://[2000::1]/temp | 469 | | 2000::1 | | " | 470 | | | | | 471 | 'http://' | 10 bytes: | (none) | "http://[DA::123:45]:6 | 472 | | ::123:45 + 616 | | 16" | 473 | | | | | 474 | 'http:///te | 18 bytes: | (none) | "http://[2000::1]:616/ | 475 | mp' | 2000::1 + 616 | | temp" | 476 +-------------+------------------+---------+------------------------+ 478 A.4. Length-aware number encoding (o256) 480 The number encoding defined in Appendix A of [I-D.ietf-core-coap] has 481 one significant flaw: Every number has an infinite number of 482 representations, which can be derived by adding leading zero bytes. 483 This runs against the principle of minimizing unnecessary choice. 484 The resulting uncertainty in encoding ultimately leads to unnecessary 485 interoperability failures. (It also wastes a small fraction of the 486 encoding space, i.e., it wastes bytes.) 488 We could solve the first, but not the second, by outlawing leading 489 zeroes, but then we have to cope with error cases caused by illegal 490 values, another source of interoperability problems. 492 The number encoding "o256" defined in this section avoids this flaw. 493 The suggestion is not to replace CoAP's "uint" encoding wholesale 494 (CoAP is already too widely implemented for such a change), but to 495 consider this format for new options. 497 The basic requirements for such an encoding are: 499 o numbers are encoded as a sequence of zero or more bytes 501 o each number has exactly one encoding 502 o for a < b, encoding-size(a) <= encoding-size(b) -- i.e., with 503 larger numbers, the encoding only gets larger, never smaller 504 again. 506 o within each encoding size (0 bytes, 1 byte, etc.), lexicographical 507 ordering of the bytes is the same as numeric ordering 509 Obviously, there is only one encoding that satisfies all these 510 requirements. As illustrated by Figure 1, this is unambiguously 511 derived by 513 1. enumerating all possible byte sequences, ordered by length and 514 within the same length in lexicographic ordering, and, 516 2. assigning sequential cardinals. 518 0x'' -> 0 519 0x'00' -> 1 520 0x'01' -> 2 521 0x'02' -> 3 522 ... 523 0x'fe' -> 255 524 0x'ff' -> 256 525 0x'0000' -> 257 526 0x'0001' -> 258 527 ... 528 0x'fefd' -> 65534 529 0x'fefe' -> 65535 530 0x'feff' -> 65536 531 ... 532 0x'ffff' -> 65792 533 0x'000000' -> 65793 534 0x'000001' -> 65794 536 Figure 1: Enumerating byte sequences by length and then lexicographic 537 order 539 This results in an exceedingly simple algorithm: each byte is 540 interpreted in the base-256 place-value system, but stands for a 541 number between 1 and 256 instead of 0 to 255. We therefore call this 542 encoding "o256" (one-to-256). 0 is always encoded in zero bytes; 1 to 543 256 is one byte, 257 (0x101) to 65792 (0x10100) is two bytes, 65793 544 (0x10101) to 16843008 (0x1010100) is three bytes, etc. 546 To further illustrate the algorithmic simplicity, pseudocode for 547 encoding and decoding is given in Figure 2 and Figure 3, respectively 548 (in the encoder, "prepend" stands for adding a byte at the _leading_ 549 edge, the requirement for which is a result of the network byte 550 order). Note that this differs only in a single subtraction/addition 551 (resp.) of one from the canonical algorithm for Appendix A uints. 553 while num > 0 554 num -= 1 555 prepend(num & 0xFF) 556 num >>= 8 557 end 559 Figure 2: o256 encoder (pseudocode) 561 num = 0 562 each_byte do |b| 563 num <<= 8 564 num += b + 1 565 end 567 Figure 3: o256 decoder (pseudocode) 569 On a more philosophical note, it can be observed that o256 solves the 570 inverse problem of Self-Delimiting Numeric Values (SDNV) [RFC6256]: 571 SDNV encodes variable-length numbers together with their length 572 (allowing decoding without knowing their length in advance, deriving 573 delimiting information from the number encoding). o256 encodes 574 variable-length numbers when there is a way to separately convey the 575 length (as in CoAP options), encoding (and later deriving) a small 576 part of the numeric value into/from that size information. 578 A.5. SMS encoding 580 For use in SMS applications, CoAP messages can be transferred using 581 SMS binary mode. However, there is operational experience showing 582 that some environments cannot successfully send a binary mode SMS. 584 For transferring SMS in character mode (7-bit characters), base64- 585 encoding [RFC4648] is an obvious choice. 3 bytes of message (24 bits) 586 turn into 4 characters, which cosume 28 bits. The overall overhead 587 is approximately 17 %; the maximum message size is 120 bytes (160 SMS 588 characters). 590 If a more compact encoding is desired, base85 encoding can be 591 employed (however, probably not the version defined in [RFC1924] -- 592 instead, the version used in tools such as btoa and PDF should be 593 chosen). However, this requires division operations. Also, the 594 base85 character set includes several characters that cannot be 595 transferred in a single 7-bit unit in SMS and/or are known to cause 596 operational problems. A modified base85 character set can be defined 597 to solve the latter problem. 4 bytes of message (32 bits) turn into 5 598 characters, which consume 35 bits. The overall overhead is 599 approximately 9.3 %; the resulting maximum message size is 128 bytes 600 (160 SMS characters). 602 Base64 and base85 do not make use of the fact that much CoAP data 603 will be ASCII-based. Therefore, we define the following experimental 604 SMS encoding. 606 A.5.1. ASCII-optimized SMS encoding 608 Not all 128 theoretically possible SMS characters are operationally 609 free of problems. We therefore define: 611 Shunned code characters: @ sign, as it maps to 0x00 613 LF and CR signs (0x0A, 0x0D) 615 uppercase C cedilla (0x09), as it is often mistranslated in 616 gateways 618 ESC (0x1B), as it is used in certain character combinations only 620 Some ASCII characters cannot be transferred in the base SMS character 621 set, as their code positions are taken by non-ASCII characters. 622 These are simply encoded with their ASCII code positions, e.g., an 623 underscore becomes a section mark (even though underscore has a 624 different code position in the SMS character set). 626 Equivalently translated input bytes: $, @, [, \, ], ^, _, `, {, |, 627 }, ~, DEL 629 In other words, bytes 0x20 to 0x7F are encoded into the same code 630 positions in the 7-bit character set. 632 Out of the remaining code characters, the following SMS characters 633 are available for encoding: 635 Non-equivalently translated (NET) code characters: 0x01 to 0x08, (8 636 characters) 638 0x0B, 0x0C, (2 characters) 640 0x0E to 0x1A, (13 characters) 641 0x1C to 0x1F, (4 characters) 643 Of the 27 NET code characters, 18 are taken as prefix characters (see 644 below), and 8 are defined as directly translated characters: 646 Directly translated bytes: Equivalently translated input bytes are 647 represented as themselves 649 0x00 to 0x07 are represented as 0x01 to 0x08 651 This leaves 0x08 to 0x1F and 0x80 to 0xFF. Of these, the bytes 0x80 652 to 0x87 and 0xA0 to 0xFF are represented as the bytes 0x00 to 0x07 653 (represented by characters 0x01 to 0x08) and 0x20 to 0x7F, with a 654 prefix of 1 (see below). The characters 0x08 to 0x1F are represented 655 as the characters 0x28 to 0x3F with a prefix of 2 (see below). The 656 characters 0x88 to 0x9F are represented as the characters 0x48 to 657 0x5F with a prefix of 2 (see below). (Characters 0x01 to 0x08, 0x20 658 to 0x27, 0x40 to 0x47, and 0x60 to 0x7f with a prefix of 2 are 659 reserved for future extensions, which could be used for some 660 backreferencing or run-length compression.) 662 Bytes that do not need a prefix (directly translated bytes) are sent 663 as is. Any byte that does need a prefix (i.e., 1 or 2) is preceded 664 by a prefix character, which provides a prefix for this and the 665 following two bytes as follows: 667 +------+-----+---+------+-----+ 668 | 0x0B | 100 | | 0x15 | 200 | 669 +------+-----+---+------+-----+ 670 | 0x0C | 101 | | 0x16 | 201 | 671 | | | | | | 672 | 0x0E | 102 | | 0x17 | 202 | 673 | | | | | | 674 | 0x0F | 110 | | 0x18 | 210 | 675 | | | | | | 676 | 0x10 | 111 | | 0x19 | 211 | 677 | | | | | | 678 | 0x11 | 112 | | 0x1A | 212 | 679 | | | | | | 680 | 0x12 | 120 | | 0x1C | 220 | 681 | | | | | | 682 | 0x13 | 121 | | 0x1D | 221 | 683 | | | | | | 684 | 0x14 | 122 | | 0x1E | 222 | 685 +------+-----+---+------+-----+ 687 (This leaves one non-shunned character, 0x1F, for future extension.) 688 The coding overhead of this encoding for random bytes is similar to 689 Base85, without the need for a division/multiplication. For bytes 690 that are mostly ASCII characters, the overhead can easily become 691 negative. (Conversely, for bytes that are more likely to be non- 692 ASCII than in a random sequence of bytes, the overhead becomes 693 greater.) 695 So, for instance, for the CoAP message in Figure 4: 697 ver tt code mid 698 1 ack 2.05 17033 699 content_type 40 700 token sometok 701 3c 2f 3e 3b 74 69 74 6c 65 3d 22 47 65 6e 65 72 |;title="Gener| 702 61 6c 20 49 6e 66 6f 22 3b 63 74 3d 30 2c 3c 2f |al Info";ct=0,;if="clock"| 704 3b 72 74 3d 22 54 69 63 6b 73 22 3b 74 69 74 6c |;rt="Ticks";titl| 705 65 3d 22 49 6e 74 65 72 6e 61 6c 20 43 6c 6f 63 |e="Internal Cloc| 706 6b 22 3b 63 74 3d 30 2c 3c 2f 61 73 79 6e 63 3e |k";ct=0,| 707 3b 63 74 3d 30 |;ct=0 | 709 Figure 4: CoAP response message as captured and decoded 711 The 116 byte unencoded message is shown as ASCII characters in 712 Figure 5 (\xDD stands for the byte with the hex digits DD): 714 bEB\x89\x11(\xA7sometok;title="General Info";ct=0, 715 ;if="clock";rt="Ticks";title="Internal Clock";ct=0,;ct=0 717 Figure 5: CoAP response message shown as unencoded characters 719 The equivalent SMS encoding is shown as equivalent-coded SMS 720 characters in Figure 6 (7 bits per character, \x12 is a 220 prefix 721 and \x0B is a 100 prefix, the rest is shown in equivalent encoding), 722 adding two characters of prefix overhead, for a total length of 118 723 7-bit characters or 104 (103.25 plus padding) bytes: 725 bEB\x12I1(\x0B'sometok;title="General Info";ct=0, 726 ;if="clock";rt="Ticks";title="Internal Clock";ct=0,;ct=0 728 Figure 6: CoAP response message shown as SMS-encoded characters 730 A.6. CONNECT 732 [RFC2817] defines the HTTP CONNECT method to establish a TCP tunnel 733 through a proxy so that end-to-end TLS connections can be made 734 through the proxy. Recently, a requirement for similar functionality 735 has been discussed for CoAP. This section defines a straw-man 736 CONNECT method and related methods and response codes for CoAP. 738 (IANA considerations for this section TBD.) 740 A.6.1. Requesting a Tunnel with CONNECT 742 CONNECT is allocated as a new method code in the "CoAP Method Codes" 743 registry. When a client makes a CONNECT request to an intermediary, 744 the intermediary evaluates the Uri-Host, Uri-Port, and/or the 745 authority part of the Proxy-Uri Options in a way that is defined by 746 the security policy of the intermediary. If the security policy 747 allows the allocation of a tunnel based on these parameters, the 748 method returns an empty payload and a response code of 2.30 Tunnel 749 Established. Other possible response codes include 4.03 Forbidden. 751 It may be the case that the intermediary itself can only reach the 752 requested origin server through another intermediary. In this case, 753 the first intermediary SHOULD make a CONNECT request of that next 754 intermediary, requesting a tunnel to the authority. A proxy MUST NOT 755 respond with any 2.xx status code unless it has either a direct or 756 tunnel connection established to the authority. 758 An origin server which receives a CONNECT request for itself MAY 759 respond with a 2.xx status code to indicate that a tunnel is 760 established to itself. 762 Code 2.30 "Tunnel Established" is allocated as a new response code in 763 the "CoAP Response Codes" registry. 765 A.6.2. Using a CONNECT Tunnel 767 Any successful (2.xx) response to a CONNECT request indicates that 768 the intermediary has established a tunnel to the requested host and 769 port. The tunnel is bound to the requesting end-point and the Token 770 supplied in the request (as always, the default Token is admissible). 771 The tunnel can be used by the client by making a DATAGRAM request. 773 DATAGRAM is allocated as a new method code in the "CoAP Method Codes" 774 registry. When a client makes a DATAGRAM request to an intermediary, 775 the intermediary looks up the tunnel bound to the client end-point 776 and Token supplied in the DATAGRAM request (no other Options are 777 permitted). If a tunnel is found and the intermediary's security 778 policy permits, the intermediary forwards the payload of the DATAGRAM 779 request as the UDP payload towards the host and port established for 780 the tunnel. No response is defined for this request (note that the 781 request can be given as a CON or NON request; for CON, there will be 782 an ACK on the message layer if the tunnel exists). 784 The security policy on the intermediary may restrict the allowable 785 payloads based on its security policy, possibly considering host and 786 port. An inadmissible payload SHOULD cause a 4.03 Forbidden response 787 with a diagnostic message as payload. 789 The UDP payload of any datagram received from the tunnel and admitted 790 by the security policy is forwarded to the client as the payload of a 791 2.31 "Datagram Received" response. The response does not carry any 792 Option except for Token, which identifies the tunnel towards the 793 client. 795 Code 2.31 "Datagram Received" is allocated as a new response code in 796 the "CoAP Response Codes" registry. 798 An origin server that has established a tunnel to itself processes 799 the CoAP payloads of related DATAGRAM requests as it would process an 800 incoming UDP payload, and forwards what would be outgoing UDP 801 payloads in 2.31 "Datagram Received" responses. 803 A.6.3. Closing down a CONNECT Tunnel 805 A 2.31 "Datagram Received" response may be replied to with a RST, 806 which closes down the tunnel. Similarly, the Token used in the 807 tunnel may be reused by the client for a different purpose, which 808 also closes down the tunnel. 810 Appendix B. The Museum (Things we did, but maybe not exactly this way) 812 B.1. Getting rid of artificial limitations 814 _Artificial limitations_ are limitations of a protocol or system that 815 are not rooted in limitations of actual capabilities, but in 816 arbitrary design decisions. Proper system design tries to avoid 817 artificial limitations, as these tend to cause complexity in systems 818 that need to work with these limitations. 820 E.g., the original UNIX filesystem had an artificial limitation of 821 the length of a path name component to 14 bytes. This led to a 822 cascade of workarounds in programs that manipulate file names: E.g., 823 systematically replacing a ".el" extension in a filename with a 824 ".elc" for the compiled file might exceed the limit, so all ".el" 825 files were suddenly limited to 13-byte filenames. 827 Note that, today, there still is a limitation in most file system 828 implementations, typically at 255. This just happens to be high 829 enough to rarely be of real-world concern; we will refer to this case 830 as a "painless" artificial limitation. 832 CoAP-08 had two highly recognizable artificial limitations in its 833 protocol encoding 835 o The number of options in a single message is limited to 15 max. 837 o The length of an option is limited to 270 max. 839 It has been argued that the latter limitation causes few problems, 840 just as the 255-byte path name component limitation in filenames 841 today causes few problems. Appendix B.1.1 provided a design to 842 extend this; as a precaution to future extensions of this kind, the 843 current encoding for length 270 (eight ones in the extension byte) 844 could be marked as reserved today. Since, Matthias Kovatsch has 845 proposed a simpler scheme that seems to gain favor in the WG, see 846 Appendix B.1.4. 848 The former limitation has been solved in CoAP-09. A historical 849 discussion of other approaches for going beyond 15 options is in 850 Appendix B.1.2. Appendix B.1.3 discusses implementation. 852 B.1.1. Beyond 270 bytes in a single option 854 The authors would argue that 270 as the maximum length of an option 855 is already beyond the "painless" threshold. 857 If that is not the consensus of the WG, the scheme can easily be 858 extended as in Figure 7: 860 for 15..269: 861 +---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+ 862 | Option Delta | 1 1 1 1 | Length - 15 | 863 +---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+ 864 | Option Value ... 865 +---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+ 867 for 270..65805: 868 +---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+ 869 | Option Delta | 1 1 1 1 | 1 1 1 1 1 1 1 1 | 870 +---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+ 871 | Length - 270 (in network byte order) | 872 +---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+ 873 | Option Value ... 874 +---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+ 876 Figure 7: Ridiculously Long Option Header 878 The infinite number of obvious variations on this scheme are left as 879 an exercise to the reader. 881 Again, as a precaution to future extensions, the current encoding for 882 length 270 (eight ones in the extension byte) could be marked as 883 reserved today. 885 B.1.2. Beyond 15 options 887 (This section keeps discussion that is no longer needed as we have 888 agreed to do what is documented in Appendix B.1.3). 890 The limit of 15 options is motivated by the fixed four-bit field "OC" 891 that is used for indicating the number of options in the fixed-length 892 CoAP header (Figure 8). 894 0 1 2 3 895 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 1 896 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 897 |Ver| T | OC | Code | Message ID | 898 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 899 | Options (if any) ... 900 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 901 | Payload (if any) ... 902 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 904 Figure 8: Four-byte fixed header in a CoAP Message 906 Note that there is another fixed four-bit field in CoAP: the option 907 length (Figure 9 - note that this figure is not to the same scale as 908 the previous figure): 910 0 1 2 3 4 5 6 7 911 +---+---+---+---+---+---+---+---+ 912 | Option Delta | Length | for 0..14 913 +---+---+---+---+---+---+---+---+ 914 | Option Value ... 915 +---+---+---+---+---+---+---+---+ 917 Figure 9: Short Option Header 919 Since 15 is inacceptable for a maximum option length, the all-ones 920 value (15) was taken out of the set of allowable values for the short 921 header, and a long header was introduced that allows the insertion of 922 an extension byte (Figure 10): 924 for 15..270: 925 +---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+ 926 | Option Delta | 1 1 1 1 | Length - 15 | 927 +---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+ 928 | Option Value ... 929 +---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+ 931 Figure 10: Long Option Header 933 We might want to use the same technique for the CoAP header as well. 934 There are two obvious places where the extension byte could be 935 placed: 937 1. right after the byte carrying the OC field, so the structure is 938 the same as for the option header; 940 2. right after the fixed-size CoAP header. 942 Both solutions lose the fixed-size-ness of the CoAP header. 944 Solution 1 has the disadvantage that the CoAP header is also changing 945 in structure: The extension byte is wedged between the first and the 946 second byte of the CoAP header. This is unfortunate, as the number 947 of options only comes into play when the option processing begins, so 948 it is more natural to use solution 2 (Figure 11): 950 0 1 2 3 951 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 1 952 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 953 |Ver| T | 15 | Code | Message ID | 954 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 955 | OC - 15 | Options ... 956 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 957 | Payload (if any) ... 958 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 960 Figure 11: Extended header for CoAP Messages with 15+ options 962 This would allow for up to 270 options in a CoAP message, which is 963 very likely way beyond the "painless" threshold. 965 B.1.2.1. Implementation considerations 967 For a message decoder, this extension creates relatively little pain, 968 as the number of options only becomes interesting when the encoding 969 turns to the options part of the message, which is then simply lead 970 in by the extension byte if the four-bit field is 15. 972 For a message encoder, this extension is not so rosy. If the encoder 973 is constructing the message serially, it may not know in advance 974 whether the number of options will exceed 14. None of the following 975 implementation strategies is particularly savory, but all of them do 976 work: 978 1. Encode the options serially under the assumption that the number 979 of options will be 14 or less. When the 15th option needs to be 980 encoded, abort the option encoding, and restart it from scratch 981 one byte further to the left. 983 2. Similar to 1, except that the bytes already encoded are all moved 984 one byte to right, the extension byte is inserted, and the option 985 encoding process is continued. 987 3. The encoder always leaves space for the extension byte (at least 988 if it can't prove the number will be less thatn 14). If the 989 extension byte is not needed, an Option 0 with length 0 is 990 encoded instead (i.e., one byte is wasted - this option is 991 elective and will be ignored by the receiver). 993 As a minimum, to enable strategy 3, the option 0 should be reserved 994 at least for the case of length=0. 996 B.1.2.2. What should we do now? 998 As a minimum proposal for the next version of CoAP, the value 15 for 999 OC should be marked as reserved today. 1001 B.1.2.3. Alternatives 1003 One alternative that has been discussed previously is to have an 1004 "Options" Option, which allows the carriage of multiple options in 1005 the belly of a single one. This could also be used to carry more 1006 than 15 options. However: 1008 o The conditional introduction of an Options option has 1009 implementation considerations that are likely to be more severe 1010 than the ones listed above; 1012 o since 270 bytes may not be enough for the encoding of _all_ 1013 options, the "Options" option would need to be repeatable. This 1014 creates many different ways to encode the same message, leading to 1015 combinatorial explosion in test cases for ensuring 1016 interoperability. 1018 B.1.2.4. Alternative: Going to a delimiter model 1020 Another alternative is to spend the additional byte not as an 1021 extended count, but as an option terminator. 1023 B.1.3. Implementing the option delimiter for 15 or more options 1025 Implementation note: As can be seen from the proof of concept code 1026 in Figure 12, the actual implementation cost for a decoder is 1027 around 4 lines of code (or about 8-10 machine code instructions). 1029 while numopt > 0 1030 nextbyte = ... get next byte 1032 if numopt == 15 # new 1033 break if nextbyte == 0xF0 # new 1034 else # new 1035 numopt -= 1 1036 end # new 1038 ... decode the delta and length from nextbyte and handle them 1039 end 1041 Figure 12: Implementing the Option Terminator 1043 Similarly, creating the option terminator needs about four more lines 1044 (not marked "old" in the C code in Figure 13). 1046 b0 = 0x40 + (tt << 4); /* old */ 1047 buffer[0] = b0 + 15; /* guess first byte */ 1049 .... encode options .... /* old */ 1051 if (option_count >= 15 || first_fragment_already_shipped) 1052 buffer[pos++] = 0xF0; /* use delimiter */ 1053 else /* save a byte: */ 1054 buffer[0] = b0 + option_count; /* old: backpatch */ 1056 Figure 13: Creating the Option Terminator 1058 B.1.4. Option Length encoding beyond 270 bytes 1060 For option lengths beyond 270 bytes, we reserve the value 255 of an 1061 extension byte to mean "add 255, read another extension byte" 1062 Figure 14. While this causes the length of the option header to grow 1063 linearly with the size of the option value, only 0.4 % of that size 1064 is used. With a focus on short options, this encoding is justified. 1066 for 15..269: 1067 +---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+ 1068 | Option Delta | 1 1 1 1 | Length - 15 | 1069 +---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+ 1070 | Option Value ... 1071 +---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+ 1073 for 270..524: 1074 +---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+ 1075 | Option Delta | 1 1 1 1 | 1 1 1 1 1 1 1 1 | 1076 +---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+ 1077 | Length - 270 | Option Value ... 1078 +---+---+---+---+---+---+---+---+ 1079 | 1080 +---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+ 1082 for 525..779: 1083 +---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+ 1084 | Option Delta | 1 1 1 1 | 1 1 1 1 1 1 1 1 | 1085 +---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+ 1086 | 1 1 1 1 1 1 1 1 | Length - 525 | 1087 +---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+ 1088 | Option Value ... 1089 +---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+ 1091 for 780..1034: 1092 +---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+ 1093 | Option Delta | 1 1 1 1 | 1 1 1 1 1 1 1 1 | 1094 +---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+ 1095 | 1 1 1 1 1 1 1 1 | 1 1 1 1 1 1 1 1 | 1096 +---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+ 1097 | Length - 780 | Option Value ... 1098 +---+---+---+---+---+---+---+---+ 1099 | 1100 +---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+ 1102 Figure 14: Options beyond 270 bytes 1104 Options that are longer than 1034 bytes MUST NOT be sent; an option 1105 that has 255 (all one bits) in the field called "Length - 780" MUST 1106 be rejected upon reception as an invalid option. 1108 In the process, the maximum length of all options that are currently 1109 set at 270 should now be set to a carefully chosen value. With the 1110 purely encoding-based limit gone, Uri-Proxy should now be restored to 1111 be a non-repeatable option. 1113 A first proposal for a new set of per-option length restrictions 1114 follows: 1116 +--------+---------------------+-----+------+--------+--------+ 1117 | number | name | min | max | type | repeat | 1118 +--------+---------------------+-----+------+--------+--------+ 1119 | 1 | content_type | 0 | 2 | uint | - | 1120 | | | | | | | 1121 | 2 | max_age | 0 | 4 | uint | - | 1122 | | | | | | | 1123 | 3 | proxy_uri | 1 | 1023 | string | - | 1124 | | | | | | | 1125 | 4 | etag | 1 | 8 | opaque | yes | 1126 | | | | | | | 1127 | 5 | uri_host | 1 | 255 | string | - | 1128 | | | | | | | 1129 | 6 | location_path | 0 | 255 | string | yes | 1130 | | | | | | | 1131 | 7 | uri_port | 0 | 2 | uint | - | 1132 | | | | | | | 1133 | 8 | location_query | 0 | 255 | string | yes | 1134 | | | | | | | 1135 | 9 | uri_path | 0 | 255 | string | yes | 1136 | | | | | | | 1137 | 10 | observe | 0 | 2 | uint | - | 1138 | | | | | | | 1139 | 11 | token | 1 | 8 | opaque | - | 1140 | | | | | | | 1141 | 12 | accept | 0 | 2 | uint | yes | 1142 | | | | | | | 1143 | 13 | if_match | 0 | 8 | opaque | yes | 1144 | | | | | | | 1145 | 14 | registered_elective | 1 | 1023 | opaque | yes | 1146 | | | | | | | 1147 | 15 | uri_query | 1 | 255 | string | yes | 1148 | | | | | | | 1149 | 17 | block2 | 0 | 3 | uint | - | 1150 | | | | | | | 1151 | 18 | size | 0 | 4 | uint | - | 1152 | | | | | | | 1153 | 19 | block1 | 0 | 3 | uint | - | 1154 | | | | | | | 1155 | 21 | if_none_match | 0 | 0 | empty | - | 1156 | | | | | | | 1157 | 25 | registered_critical | 1 | 1023 | opaque | yes | 1158 +--------+---------------------+-----+------+--------+--------+ 1160 (Option 14 with a length of 0 is a fencepost only.) 1162 B.2. Registered Option 1164 CoAP's option encoding is highly efficient, but works best with small 1165 option numbers that do not require much fenceposting. The CoAP 1166 Option Number Registry therefore has a relatively heavyweight 1167 registration requirement: "IETF Review" as described in [RFC5226]. 1169 However, there is also considerable benefit in a much looser registry 1170 policy, enabling a first-come-first-served policy for a relatively 1171 large option number space. 1173 Here, we discuss two solutions that enable such a registry. One is 1174 to define a separate mechanism for registered options, discussed in 1175 Appendix B.2.1. Alternatively, we could make it easier to use a 1176 larger main option number space, discussed in Appendix B.2.2. 1178 B.2.1. A Separate Suboption Number Space 1180 This alternative defines a separate space of suboption numbers, with 1181 an expert review [RFC5226] (or even first-come-first-served) 1182 registration policy. If expert review is selected for this registry, 1183 it would be with a relatively loose policy delegated to the expert. 1184 This draft proposes leaving the registered suboption numbers 0-127 to 1185 expert review with a policy that mainly focuses on the availability 1186 of a specification, and 128-16383 for first-come-first-served where 1187 essentially only a name is defined. 1189 The "registered" options are used in conjunction with this suboption 1190 number registry. They use two normal CoAP option numbers, one for 1191 options with elective semantics (Registered-Elective) and one for 1192 options with critical semantics (Registered-Critical). The suboption 1193 numbers are not separate, i.e. one registered suboption number might 1194 have some elective semantics and some other critical semantics (e.g., 1195 for the request and the response leg of an exchange). The option 1196 value starts with an SDNV [RFC6256] of the registered suboption 1197 number. (Note that there is no need for an implementation to 1198 understand SDNVs, it can treat the prefixes as opaque. One could 1199 consider the SDNVs as a suboption prefix allocation guideline for 1200 IANA as opposed to a number encoding.) 1202 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1203 |1 0 0 0 0 0 0 1|0 1 1 1 0 0 1 1| value... | 1204 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1205 \___SDNV of registered number___/ 1207 Figure 15: Example option value for registered option 1209 Note that a Registered Option cannot be empty, because there would be 1210 no space for the SDNV. Also, the empty option 14 is reserved for 1211 fenceposting ([I-D.ietf-core-coap], section 3.2). (Obviously, once a 1212 Registered-Elective Option is in use, there is never a need for a 1213 fence-post for option number 14.) 1215 The Registered-Elective and Registered-Critical Options are 1216 repeatable. 1218 +-----+----------+---------------------+---------+--------+---------+ 1219 | No. | C/E | Name | Format | Length | Default | 1220 +-----+----------+---------------------+---------+--------+---------+ 1221 | 14 | Elective | Registered-Elective | (see | 1-1023 | (none) | 1222 | | | | above) | B | | 1223 | | | | | | | 1224 | 25 | Critical | Registered-Critical | (see | 1-1023 | (none) | 1225 | | | | above) | B | | 1226 +-----+----------+---------------------+---------+--------+---------+ 1228 This solves CoRE issue #214 [CoRE214]. (How many options we need 1229 will depend on the resolution of #241 [CoRE241].) 1231 B.2.2. Opening Up the Option Number Space 1233 The disadvantage of the registered-... options is that there is a 1234 significant syntactic difference between options making use of this 1235 space and the usual standard options. This creates a problem not 1236 unlike that decried in [RFC6648]. 1238 The alternative discussed in this section reduces the distance by 1239 opening up the main Option number space instead. 1241 There is still a significant incentive to use low-numbered Options. 1242 However, the proposal reduces the penalty for using a high-numbered 1243 Option to two or three bytes. More importantly, using a cluster of 1244 related high-numbered options only carries a total penalty of two or 1245 three bytes. 1247 The main reason high-numbered options are expensive to use and thus 1248 the total space is relatively limited is that the option delta 1249 mechanism only allows increasing the current option number by up to 1250 14 per one-byte fencepost. To use, e.g., Option number 1234 together 1251 with the usual set of low-numbered Options, one needs to insert 88 1252 fence-post bytes. This is prohibitive. 1254 Enabling first-come-first-served probably requires easily addressing 1255 a 16-bit option number space, with some potential increase later in 1256 the lifetime of the protocol (say, 10 to 15 years from now). 1258 To enable the use of large option numbers, one needs a way to advance 1259 the Option number in bigger steps than possible by the Option Delta. 1260 So we propose a new construct, the Long Jump construct, to move the 1261 Option number forward. 1263 B.2.2.1. Long Jump construct 1265 The following construct can occur in front of any Option: 1267 0 1 2 3 4 5 6 7 1268 +---+---+---+---+---+---+---+---+ 1269 | 1 1 1 1 | 0 0 0 1 | 0xf1 (Delta = 15) 1270 +---+---+---+---+---+---+---+---+ 1272 0 1 2 3 4 5 6 7 1273 +---+---+---+---+---+---+---+---+ 1274 | 1 1 1 1 | 0 0 1 0 | 0xf2 1275 +---+---+---+---+---+---+---+---+ 1276 | Long Jump Value | (Delta/8)-2 1277 +---+---+---+---+---+---+---+---+ 1279 0 1 2 3 4 5 6 7 1280 +---+---+---+---+---+---+---+---+ 1281 | 1 1 1 1 | 0 0 1 1 | 0xf3 1282 +---+---+---+---+---+---+---+---+ 1283 | Long Jump Value, MSB | 1284 +---+---+---+---+---+---+---+---+ (Delta/8)-258 1285 | Long Jump Value, LSB | 1286 +---+---+---+---+---+---+---+---+ 1288 Figure 16: Long Jump Format 1290 This construct is not by itself an Option. It can occur in front of 1291 any Option to increase the current Option number that then goes into 1292 its Option number calculation. The increase is done in multiples of 1293 eight. More specifically, the actual addition to the current Option 1294 number is computed as follows: 1296 Delta = ((Long Jump Value) + N) * 8 1298 where N is 2 for the one-byte version and N is 258 for the two-byte 1299 version. 1301 A Long Jump MUST be followed by an actual Option, i.e., it MUST NOT 1302 be followed by another Long Jump or an end-of-options indicator. A 1303 message violating this MUST be rejected as malformed. 1305 Long Jumps do NOT count as Options in the Option Count field of the 1306 header (i.e., they cannot by themselves end the Option sequence). 1308 B.2.2.2. Discussion 1310 Adding a mechanism at this late stage creates concerns of backwards 1311 compatibility. A message sender never needs to implement long-jumps 1312 unless it wants to make use of a high-numbered option. So this 1313 mechanism can be added once a high-numbered option is added. A 1314 message receiver, though, would more or less unconditionally have to 1315 implement the mechanism, leading to unconditional additional 1316 complexity. There are good reasons to minimize this, as follows: 1318 o The increase in multiples of eight allows looking at an option and 1319 finding out whether it is critical or not even if the Long Jump 1320 value has just been skipped (as opposed to having been processed 1321 fully). (It also allows accessing up to approximately 2048 1322 options with a two-byte Long Jump.) This allows a basic 1323 implementation that does not implement any high-numbered options 1324 to simply ignore long jumps and any elective options behind them, 1325 while still properly reacting to critical options. 1327 o There is probably a good reason to disallow long-jumps that lead 1328 to an option number of 42 and less, enabling simple receivers to 1329 do the above simplification. 1331 o It might seem obvious to remove the fenceposting mechanism 1332 altogether in favor of long jumps. This is not advisable: 1333 Fenceposting already has zero implementation effort at the 1334 receiver, and the overhead at the sender is very limited (it is 1335 just a third kind of jump, at one byte per jump). Beyond 42, 1336 senders can ignore the existence of fenceposts if they want 1337 (possibly obviating the need for more complex base-14 arithmetic). 1339 There is no need for a finer granularity than 8, as the Option 1340 construct following can also specify a Delta of 0..14. (A 1341 granularity of 16 will require additional fenceposting where an 1342 option delta of 15 would happen to be required otherwise, which we 1343 have reserved. It can be argued that 16 is still the better choice, 1344 as fenceposting is already in the code path.) 1346 The Long Jump construct takes 0xf1 and 0xf2 from the space available 1347 for initial bytes of Options. (Note that we previously took 0xf0 to 1348 indicate end-of-options for OC=15.) 1350 Varying N with the length as defined above makes it unambiguous 1351 whether a one- or two-byte Long Jump is to be used. Setting N=2 for 1352 the one-byte version makes it clear that a Delta of 8 is to be 1353 handled the usual way (i.e., by Option Delta itself and/or 1354 fenceposting). If the delta is not small and not 7 modulo 8, there 1355 is still a choice between using the smaller multiple of 8 and a 1356 larger Delta in the actual Option or v.v., this biases the choice 1357 towards a larger Long Jump and a smaller following Delta, which is 1358 also easier to implement as it reduces the number of choice points. 1360 B.2.2.3. Example 1362 The following sequence of bytes would encode a Uri-Path Option of 1363 "foo" followed by Options 1357 (value "bar") and 1360 (value "baz"): 1365 93 65 6f 6f Option 9 (0 + 9, "foo") 1366 f1 a6 Long Jump by 1344 1367 43 62 61 72 Option 1357 (9 + 1344 + 4, "bar") 1368 33 62 61 7a Option 1360 (1357 + 3, "baz") 1370 Figure 17: Example using a Long Jump construct 1372 where f1 a6 is the long jump forward by (0xa6+2)*8=1344 option 1373 numbers. The total option count (OC) for the CoAP header is 3. Note 1374 that even if f1 a6 is skipped, the 1357 (which then appears as an 1375 Option number 13) is clearly visible as Critical. 1377 B.2.2.4. IANA considerations 1379 With the scheme proposed above, we could have three tiers of Option 1380 Numbers: 1382 +---------------+-------------------------+ 1383 | Option Number | Policy [RFC5226] | 1384 +---------------+-------------------------+ 1385 | 0..255 | Standards Action | 1386 | | | 1387 | 256..2047 | Designated Expert | 1388 | | | 1389 | 2048..65535 | First Come First Served | 1390 +---------------+-------------------------+ 1392 For the inventor of a new option, this would provide a small 1393 incentive to go through the designated expert for some minimal cross- 1394 checking in order to be able to use the two-byte long-jump. 1396 This draft adds option numbers to Table 2 of [I-D.ietf-core-coap]: 1398 +--------+---------------------+-----------+ 1399 | Number | Name | Reference | 1400 +--------+---------------------+-----------+ 1401 | 14 | Registered-Elective | [RFCXXXX] | 1402 | | | | 1403 | 25 | Registered-Critical | [RFCXXXX] | 1404 +--------+---------------------+-----------+ 1406 Table 1: New CoAP Option Numbers 1408 This draft adds a suboption registry, initially empty. 1410 +------------+-----------------------------+-----------+ 1411 | Number | Name | Reference | 1412 +------------+-----------------------------+-----------+ 1413 | 0..127 | (allocate on export review) | [RFCXXXX] | 1414 | | | | 1415 | 128..16383 | (allocate fcfs) | [RFCXXXX] | 1416 +------------+-----------------------------+-----------+ 1418 Table 2: CoAP Suboption Numbers 1420 B.3. Enabling Protocol Evolution 1422 To enable a protocol to evolve, it is critical that new capabilities 1423 can be introduced without requiring changes in components that don't 1424 really care about the capability. One such probem is exhibited by 1425 CoAP options: If a proxy does not understand an elective option in a 1426 request, it will not be able to forward it to the origin server, 1427 rendering the new option ineffectual. Worse, if a proxy does not 1428 understand a critical option in a request, it will not be able to 1429 operate on the request, rendering the new option damaging. 1431 As a conclusion to the Ticket #230 discussion in the June 4th interim 1432 call, we decided to solve the identification of options that a proxy 1433 can safely forward even if not understood (previously called Proxy- 1434 Elective). 1436 The proposal is to encode this information in the option number, just 1437 like the way the information that an option is critical is encoded 1438 now. This leads to two bits with semantics: the lowest bit continues 1439 to be the critical bit, and the next higher bit is now the "unsafe" 1440 bit (i.e., this option is not safe to forward unless understood by 1441 the proxy). 1443 Another consideration (for options that are not unsafe to forward) is 1444 whether the option should serve as a cache key in a request. HTTP 1445 has a vary header that indicates in the response which header fields 1446 were considered by the origin server to be cache keys. In order to 1447 avoid this complexity, we should be able to indicate this information 1448 right in the option number. However, reserving another bit is 1449 wasteful, in particular as there are few safe-to-forward options that 1450 are not cache-keys. 1452 Therefore, we propose the following bit allocation in an option 1453 number: 1455 xxx nnn UC 1457 Figure 18 1459 (where xxx is a variable length prefix, as option numbers are not 1460 bounded upwards). UC is the unsafe and critical bits. For U=0 only, 1461 if nnn is equal to 111 binary, the option does not serve as a cache 1462 key (for U=1, the proxy has to know the option to act on it, so there 1463 is no point in indicating whether it is a cache key). There is no 1464 semantic meaning of xxx. 1466 Note that clients and servers are generally not interested in this 1467 information. A proxy may use an equivalent of the following C code 1468 to derive the characteristics of an option number "onum": 1470 Critical = (onum & 1); 1471 UnSafe = (onum & 2); 1472 NoCache = ((onum & 0x1e) == 0x1c); 1474 Figure 19 1476 Discussion: This requires a renumbering of all options. 1478 This renumbering may also be considered as an opportunity to make 1479 the numbering straight again shortly before nailing down the 1480 protocol 1482 In particular, Content-Type is now probably better considered to 1483 be elective. 1485 B.3.1. Potential new option number allocation 1487 We want to give one example for a revised allocation of option 1488 numbers. Option numbers are given as decimal numbers, one each for 1489 xxx, nnn, and UC, with the UC values as follows 1490 +-----------+------------+------------------------------------+ 1491 | UC binary | UC decimal | meaning | 1492 +-----------+------------+------------------------------------+ 1493 | 00 | 0 | (safe, elective, 111=no-cache-key) | 1494 | | | | 1495 | 01 | 1 | (safe, critical, 111=no-cache-key) | 1496 | | | | 1497 | 10 | 2 | (unsafe, elective) | 1498 | | | | 1499 | 11 | 3 | (unsafe, critical) | 1500 +-----------+------------+------------------------------------+ 1502 The table is: 1504 +-----+---------+-------+-------------------+-----------------------+ 1505 | New | xx nnn | Old | Name | Comment | 1506 | | UC | | | | 1507 +-----+---------+-------+-------------------+-----------------------+ 1508 | 4 | 0 1 0 | 1 | Content-Type | category change | 1509 | | | | | (elective) | 1510 | | | | | | 1511 | 8 | 0 2 0 | 4 | ETag | | 1512 | | | | | | 1513 | 12 | 0 3 0 | 12 | Accept | | 1514 | | | | | | 1515 | 16 | 0 4 0 | 6 | Location-Path | | 1516 | | | | | | 1517 | 20 | 0 5 0 | 8 | Location-Query | | 1518 | | | | | | 1519 | 24 | 0 6 0 | - | (unused) | | 1520 | | | | | | 1521 | 28 | 0 7 0 | 18 | Size | needs nnn=111 | 1522 | | | | | | 1523 | 32 | 1 0 0 | 20/22 | Patience | | 1524 | | | | | | 1525 | 64 | 2 x 0 | - | Location-reserved | (nnn = 0..3, 4 | 1526 | | | | | reserved numbers) | 1527 | | | | | | 1528 | 1 | 0 0 1 | 13 | If-Match | | 1529 | | | | | | 1530 | 5 | 0 1 1 | 21 | If-None-Match | | 1531 | | | | | | 1532 | 2 | 0 0 2 | 2 | Max-Age | | 1533 | | | | | | 1534 | 6 | 0 1 2 | 10 | Observe | | 1535 | | | | | | 1536 | 10 | 0 2 2 | xx | Observe-2 | | 1537 | | | | | | 1538 | 14 | 0 3 2 | xx | (unused) | was fencepost | 1539 | | | | | | 1540 | 3 | 0 0 3 | 3 | Proxy-Uri | | 1541 | | | | | | 1542 | 7 | 0 1 3 | 5 | Uri-Host | | 1543 | | | | | | 1544 | 11 | 0 2 3 | 7 | Uri-Port | | 1545 | | | | | | 1546 | 15 | 0 3 3 | 9 | Uri-Path | | 1547 | | | | | | 1548 | 19 | 0 4 3 | 15 | Uri-Query | | 1549 | | | | | | 1550 | 23 | 0 5 3 | 11 | Token | | 1551 | | | | | | 1552 | 27 | 0 6 3 | 17 | Block2 | | 1553 | | | | | | 1554 | 31 | 0 7 3 | 19 | Block1 | yes, we can use | 1555 | | | | | nnn=111 with U=1 | 1556 +-----+---------+-------+-------------------+-----------------------+ 1558 B.4. Patience, Leisure, and Pledge 1560 A number of options might be useful for controlling the timing of 1561 interactions. 1563 (This section also addresses core-coap ticket #177.) 1565 B.4.1. Patience 1567 A client may have a limited time period in which it can actually make 1568 use of the response for a request. Using the Patience option, it can 1569 provide an (elective) indication how much time it is willing to wait 1570 for the response from the server, giving the server license to ignore 1571 or reject the request if it cannot fulfill it in this period. 1573 If the server knows early that it cannot fulfill the request in the 1574 time requested, it MAY indicate this with a 5.04 "Timeout" response. 1575 For non-safe methods (such as PUT, POST, DELETE), the server SHOULD 1576 indicate whether it has fulfilled the request by either responding 1577 with 5.04 "Timeout" (and not further processing the request) or by 1578 processing the request normally. 1580 Note that the value of the Patience option should be chosen such that 1581 the client will be able to make use of the result even in the 1582 presence of the expected network delays for the request and the 1583 response. Similarly, when a proxy receives a request with a Patience 1584 option and cannot fulfill that request from its cache, it may want to 1585 adjust the value of the option before forwarding it to an upstream 1586 server. 1588 (TBD: The various cases that arise when combining Patience with 1589 Observe.) 1591 The Patience option is elective. Hence, a client MUST be prepared to 1592 receive a normal response even after the chosen Patience period (plus 1593 an allowance for network delays) has elapsed. 1595 B.4.2. Leisure 1597 Servers generally will compute an internal value that we will call 1598 Leisure, which indicates the period of time that will be used for 1599 responding to a request. A Patience option, if present, can be used 1600 as an upper bound for the Leisure. Leisure may be non-zero for 1601 congestion control reasons, in particular for responses to multicast 1602 requests. For these, the server should have a group size estimate G, 1603 a target rate R (which both should be chosen conservatively) and an 1604 estimated response size S; a rough lower bound for Leisure can then 1605 be computed as follows: 1607 lb_Leisure = S * G / R 1609 Figure 20: Computing a lower bound for the Leisure 1611 E.g., for a multicast request with link-local scope on an 2.4 GHz 1612 IEEE 802.15.4 (6LoWPAN) network, G could be (relatively 1613 conservatively) set to 100, S to 100 bytes, and the target rate to 8 1614 kbit/s = 1 kB/s. The resulting lower bound for the Leisure is 10 1615 seconds. 1617 To avoid response implosion, responses to multicast requests SHOULD 1618 be dithered within a Leisure period chosen by the server to fall 1619 between these two bounds. 1621 Currently, we don't foresee a need to signal a value for Leisure from 1622 client to server (beyond the signalling provided by Patience) or from 1623 server to client, but an appropriate Option might be added later. 1625 B.4.3. Pledge 1627 In a basic observation relationship [I-D.ietf-core-observe], the 1628 server makes a pledge to keep the client in the observation 1629 relationship for a resource at least until the max-age for the 1630 resource is reached. 1632 To save the client some effort in re-establishing observation 1633 relationships each time max-age is reached, the server MAY want to 1634 extend its pledge beyond the end of max-age by signalling in a 1635 response/notification an additional time period using the Pledge 1636 Option, in parallel to the Observe Option. 1638 The Pledge Option MUST NOT be used unless the server can make a 1639 reasonable promise not to lose the observation relationship in this 1640 time frame. 1642 Currently, we don't foresee a need to signal a value for Pledge from 1643 client to server, but an appropriate behavior might be added later 1644 for this option when sent in a request. 1646 B.4.4. Option Formats 1648 +-----+----------+----------+-----------------+--------+---------+ 1649 | No. | C/E | Name | Format | Length | Default | 1650 +-----+----------+----------+-----------------+--------+---------+ 1651 | 22 | Elective | Patience | Duration in mis | 1 B | (none) | 1652 | | | | | | | 1653 | 24 | Elective | Pledge | Duration in s | 1 B | 0 | 1654 +-----+----------+----------+-----------------+--------+---------+ 1656 All timing options use the Duration data type (see Appendix D.2), 1657 however Patience (and Leisure, if that ever becomes an option) uses a 1658 timebase of mibiseconds (mis = 1/1024 s) instead of seconds. (This 1659 reduces the range of the Duration from ~ 91 days to 128 minutes.) 1661 Implementation note: As there are no strong accuracy requirements on 1662 the clocks employed, making use of any existing time base of 1663 milliseconds is a valid implementation approach (2.4 % off). 1665 None of the options may be repeated. 1667 Appendix C. The Cemetery (Things we won't do) 1669 This annex documents roads that the WG decided not to take, in order 1670 to spare readers from reinventing them in vain. 1672 C.1. Example envelope option: solving #230 1674 Ticket #230 [CoRE230] points out a design flaw of 1675 [I-D.ietf-core-coap]: When we split the elective Location option of 1676 draft -01 into multiple elective options, we made it possible that an 1677 implementation might process some of these and ignore others, leading 1678 to an incorrect interpretation of the Location expressed by the 1679 server. 1681 There are several more or less savory solutions to #230. 1683 Each of the elective options that together make up the Location could 1684 be defined in such a way that it makes a requirement on the 1685 processing of the related option (essentially revoking their elective 1686 status once the option under consideration is actually processed). 1687 This falls flat as soon as another option is defined that would also 1688 become part of the Location: existing implementations would not know 1689 that the new option is also part of the cluster that is re- 1690 interpreted as critical. The potential future addition of Location- 1691 Host and Location-Port makes this a valid consideration. 1693 A better solution would be to define an elective Envelope Option 1694 called Location. Within a Location Option, the following top-level 1695 options might be allowed (now or in the future): 1697 o Uri-Host 1699 o Uri-Port 1701 o Uri-Path 1703 o Uri-Query 1705 This would unify the code for interpreting the top-level request 1706 options that indicate the request URI with the code that interprets 1707 the Location URI. 1709 The four options listed are all critical, while the envelope is 1710 elective. This gives exactly the desired semantics: If the envelope 1711 is processed at all (which is elective), the nested options are 1712 critical and all need to be processed. 1714 C.2. Example envelope option: proxy-elective options 1716 Another potential application of envelope options is motivated by the 1717 observation that new critical options might not be implemented by all 1718 proxies on the CoAP path to an origin server. So that this does not 1719 become an obstacle to introducing new critical options that are of 1720 interest only to client and origin server, the client might want to 1721 mark some critical options proxy-elective, i.e. elective for a proxy 1722 but still critical for the origin server. 1724 One way to do this would be an Envelope option, the Proxy-Elective 1725 Option. A client might bundle a number of critical options into a 1726 critical Proxy-Elective Option. A proxy that processes the message 1727 is obliged to process the envelope (or reject the message), where 1728 processing means passing on the nested options towards the origin 1729 server (preferably again within a Proxy-Elective option). It can 1730 pass on the nested options, even ones unknown to the proxy, knowing 1731 that the client is happy with proxies not processing all of them. 1733 (The assumption here is that the Proxy-Elective option becomes part 1734 of the base standard, so all but the most basic proxies would know 1735 how to handle it.) 1737 C.3. Stateful URI compression 1739 Is the approximately 25 % average saving achievable with Huffman- 1740 based URI compression schemes worth the complexity? Probably not, 1741 because much higher average savings can be achieved by introducing 1742 state. 1744 Henning Schulzrinne has proposed for a server to be able to supply a 1745 shortened URI once a resource has been requested using the full- 1746 length URI. Let's call such a shortened referent a _Temporary 1747 Resource Identifier_, _TeRI_ for short. This could be expressed by a 1748 response option as shown in Figure 21. 1750 0 1751 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 1752 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1753 | duration | TeRI... 1754 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1756 Figure 21: Option for offering a TeRI in a response 1758 The TeRI offer option indicates that the server promises to offer 1759 this resources under the TeRI given for at least the time given as 1760 the duration. Another TeRI offer can be made later to extend the 1761 duration. 1763 Once a TeRI for a URI is known (and still within its lifetime), the 1764 client can supply a TeRI instead of a URI in its requests. The same 1765 option format as an offer could be used to allow the client to 1766 indicate how long it believes the TeRI will still be valid (so that 1767 the server can decide when to update the lifetime duration). TeRIs 1768 in requests could be distinguished from URIs e.g. by using a 1769 different option number. 1771 Proposal: Add a TeRI option that can be used in CoAP requests and 1772 responses. 1774 Add a way to indicate a TeRI and its duration in a link-value. 1776 Do not add any form of stateless URI encoding. 1778 Benefits: Much higher reduction of message size than any stateless 1779 URI encoding could achieve. 1781 As the use of TeRIs is entirely optional, minimal complexity nodes 1782 can get by without implementing them. 1784 Drawbacks: Adds considerable state and complexity to the protocol. 1786 It turns out that real CoAP URIs are short enough that TeRIs are 1787 not needed. 1789 (Discuss the security implications of TeRIs.) 1791 Appendix D. Experimental Options 1793 This annex documents proposals that need significant additional 1794 discussion before they can become part of (or go back to) the main 1795 CoAP specification. They are not dead, but might die if there turns 1796 out to be no good way to solve the problem. 1798 D.1. Options indicating absolute time 1800 HTTP has a number of headers that may indicate absolute time: 1802 o "Date", defined in Section 14.18 in [RFC2616] (Section 9.3 in 1803 [I-D.ietf-httpbis-p1-messaging]), giving the absolute time a 1804 response was generated; 1806 o "Last-Modified", defined in Section 14.29 in [RFC2616], (Section 1807 6.6 in [I-D.ietf-httpbis-p4-conditional], giving the absolute time 1808 of when the origin server believes the resource representation was 1809 last modified; 1811 o "If-Modified-Since", defined in Section 14.25 in [RFC2616], 1812 "If-Unmodified-Since", defined in Section 14.28 in [RFC2616], and 1813 "If-Range", defined in Section 14.27 in [RFC2616] can be used to 1814 supply absolute time to gate a conditional request; 1816 o "Expires", defined in Section 14.21 in [RFC2616] (Section 3.3 in 1817 [I-D.ietf-httpbis-p6-cache]), giving the absolute time after which 1818 a response is considered stale. 1820 o The more obscure headers "Retry-After", defined in Section 14.37 1821 in [RFC2616], and "Warning", defined in section 14.46 in 1822 [RFC2616], also may employ absolute time. 1824 [I-D.ietf-core-coap] defines a single "Date" option, which however 1825 "indicates the creation time and date of a given resource 1826 representation", i.e., is closer to a "Last-Modified" HTTP header. 1827 HTTP's caching rules [I-D.ietf-httpbis-p6-cache] make use of both 1828 "Date" and "Last-Modified", combined with "Expires". The specific 1829 semantics required for CoAP needs further consideration. 1831 In addition to the definition of the semantics, an encoding for 1832 absolute times needs to be specified. 1834 In UNIX-related systems, it is customary to indicate absolute time as 1835 an integer number of seconds, after midnight UTC, January 1, 1970. 1836 Unless negative numbers are employed, this time format cannot 1837 represent time values prior to January 1, 1970, which probably is not 1838 required for the uses ob absolute time in CoAP. 1840 If a 32-bit integer is used and allowance is made for a sign-bit in a 1841 local implementation, the latest UTC time value that can be 1842 represented by the resulting 31 bit integer value is 03:14:07 on 1843 January 19, 2038. If the 32-bit integer is used as an unsigned 1844 value, the last date is 2106-02-07, 06:28:15. 1846 The reach can be extended by: - moving the epoch forward, e.g. by 40 1847 years (= 1262304000 seconds) to 2010-01-01. This makes it impossible 1848 to represent Last-Modified times in that past (such as could be 1849 gatewayed in from HTTP). - extending the number of bits, e.g. by one 1850 more byte, either always or as one of two formats, keeping the 32-bit 1851 variant as well. 1853 Also, the resolution can be extended by expressing time in 1854 milliseconds etc., requiring even more bits (e.g., a 48-bit unsigned 1855 integer of milliseconds would last well after year 9999.) 1857 For experiments, an experimental "Date" option is defined with the 1858 semantics of HTTP's "Last-Modified". It can carry an unsigned 1859 integer of 32, 40, or 48 bits; 32- and 40-bit integers indicate the 1860 absolute time in seconds since 1970-01-01 00:00 UTC, while 48-bit 1861 integers indicate the absolute time in milliseconds since 1970-01-01 1862 00:00 UTC. 1864 However, that option is not really that useful until there is a 1865 "If-Modified-Since" option as well. 1867 (Also: Discuss nodes without clocks.) 1869 D.2. Representing Durations 1871 Various message types used in CoAP need the representation of 1872 *durations*, i.e. of the length of a timespan. In SI units, these 1873 are measured in seconds. CoAP durations represent integer numbers of 1874 seconds, but instead of representing these numbers as integers, a 1875 more compact single-byte pseudo-floating-point (pseudo-FP) 1876 representation is used (Figure 22). 1878 0 1 2 3 4 5 6 7 1879 +---+---+---+---+---+---+---+---+ 1880 | 0... value | 1881 +---+---+---+---+---+---+---+---+ 1883 +---+---+---+---+---+---+---+---+ 1884 | 1... mantissa | exponent | 1885 +---+---+---+---+---+---+---+---+ 1886 Figure 22: Duration in (8,4) pseudo-FP representation 1888 If the high bit is clear, the entire n-bit value (including the high 1889 bit) is the decoded value. If the high bit is set, the mantissa 1890 (including the high bit, with the exponent field cleared out but 1891 still present) is shifted left by the exponent to yield the decoded 1892 value. 1894 The (n,e)-pseudo-FP format can be decoded with a single line of code 1895 (plus a couple of constant definitions), as demonstrated in 1896 Figure 23. 1898 #define N 8 1899 #define E 4 1900 #define HIBIT (1 << (N - 1)) 1901 #define EMASK ((1 << E) - 1) 1902 #define MMASK ((1 << N) - 1 - EMASK) 1904 #define DECODE_8_4(r) (r < HIBIT ? r : (r & MMASK) << (r & EMASK)) 1906 Figure 23: Decoding an (8,4) pseudo-FP value 1908 Note that a pseudo-FP encoder needs to consider rounding; different 1909 applications of durations may favor rounding up or rounding down the 1910 value encoded in the message. 1912 The highest pseudo-FP value, represented by an all-ones byte (0xFF), 1913 is reserved to indicate an indefinite duration. The next lower value 1914 (0xEF) is thus the highest representable value and is decoded as 1915 7340032 seconds, a little more than 12 weeks. 1917 D.3. Rationale 1919 Where CPU power and memory is abundant, a duration can almost always 1920 be adequately represented by a non-negative floating-point number 1921 representing that number of seconds. Historically, many APIs have 1922 also used an integer representation, which limits both the resolution 1923 (e.g., if the integer represents the duration in seconds) and often 1924 the range (integer machine types have range limits that may become 1925 relevant). UNIX's "time_t" (which is used for both absolute time and 1926 durations) originally was a signed 32-bit value of seconds, but was 1927 later complemented by an additional integer to add microsecond 1928 ("struct timeval") and then later nanosecond ("struct timespec") 1929 resolution. 1931 Three decisions need to be made for each application of the concept 1932 of duration: 1934 o the *resolution*. What rounding error is acceptable? 1936 o the *range*. What is the maximum duration that needs to be 1937 represented? 1939 o the *number of bits* that can be expended. 1941 Obviously, these decisions are interrelated. Typically, a large 1942 range needs a large number of bits, unless resolution is traded. For 1943 most applications, the actual requirement for resolution are limited 1944 for longer durations, but can be more acute for shorter durations. 1946 D.4. Pseudo-Floating Point 1948 Constrained systems typically avoid the use of floating-point (FP) 1949 values, as 1951 o simple CPUs often don't have support for floating-point datatypes 1953 o software floating-point libraries are expensive in code size and 1954 slow. 1956 In addition, floating-point datatypes used to be a significant 1957 element of market differentiation in CPU design; it has taken the 1958 industry a long time to agree on a standard floating point 1959 representation. 1961 These issues have led to protocols that try to constrain themselves 1962 to integer representation even where the ability of a floating point 1963 representation to trade range for resolution would be beneficial. 1965 The idea of introducing _pseudo-FP_ is to obtain the increased range 1966 provided by embedding an exponent, without necessarily getting stuck 1967 with hardware datatypes or inefficient software floating-point 1968 libraries. 1970 For the purposes of this draft, we define an (n,e)-pseudo-FP as a 1971 fixed-length value of n bits, e of which may be used for an exponent. 1972 Figure 22 illustrates an (8,4)-pseudo-FP value. 1974 If the high bit is clear, the entire n-bit value (including the high 1975 bit) is the decoded value. If the high bit is set, the mantissa 1976 (including the high bit, but with the exponent field cleared out) is 1977 shifted left by the exponent to yield the decoded value. 1979 The (n,e)-pseudo-FP format can be decoded with a single line of code 1980 (plus a couple of constant definition), as demonstrated in Figure 23. 1982 Only non-negative numbers can be represented by this format. It is 1983 designed to provide full integer resolution for values from 0 to 1984 2^(n-1)-1, i.e., 0 to 127 in the (8,4) case, and a mantissa of n-e 1985 bits from 2^(n-1) to (2^n-2^e)*2^(2^e-1), i.e., 128 to 7864320 in the 1986 (8,4) case. By choosing e carefully, resolution can be traded 1987 against range. 1989 Note that a pseudo-FP encoder needs to consider rounding; different 1990 applications of durations may favor rounding up or rounding down the 1991 value encoded in the message. This requires a little more than a 1992 single line of code (which is left as an exercise to the reader, as 1993 the most efficient expression depends on hardware details). 1995 D.5. A Duration Type for CoAP 1997 CoAP needs durations in a number of places. In [I-D.ietf-core-coap], 1998 durations occur in the option "Subscription-lifetime" as well as in 1999 the option "Max-age". (Note that the option "Date" is not a 2000 duration, but a point in time.) Other durations of this kind may be 2001 added later. 2003 Most durations relevant to CoAP are best expressed with a minimum 2004 resolution of one second. More detailed resolutions are unlikely to 2005 provide much benefit. 2007 The range of lifetimes and caching ages are probably best kept below 2008 the order of magnitude of months. An (8,4)-pseudo-FP has the maximum 2009 value of 7864320, which is about 91 days; this appears to be adequate 2010 for a subscription lifetime and probably even for a maximum cache 2011 age. Figure 24 shows the values that can be expressed. (If a larger 2012 range for the latter is indeed desired, an (8,5)-pseudo-FP could be 2013 used; this would last 15 milleniums, at the cost of having only 3 2014 bits of accuracy for values larger than 127 seconds.) 2016 Proposal: A single duration type is used throughout CoAP, based on 2017 an (8,4)-pseudo-FP giving a duration in seconds. 2019 Benefits: Implementations can use a single piece of code for 2020 managing all CoAP-related durations. 2022 In addition, length information never needs to be managed for 2023 durations that are embedded in other data structures: All 2024 durations are expressed by a single byte. 2026 It might be worthwhile to reserve one duration value, e.g. 0xFF, for 2027 an indefinite duration. 2029 Duration Seconds Encoded 2030 ----------- ---------- ------- 2031 00:00:00 0x00000000 0x00 2032 00:00:01 0x00000001 0x01 2033 00:00:02 0x00000002 0x02 2034 00:00:03 0x00000003 0x03 2035 00:00:04 0x00000004 0x04 2036 00:00:05 0x00000005 0x05 2037 00:00:06 0x00000006 0x06 2038 00:00:07 0x00000007 0x07 2039 00:00:08 0x00000008 0x08 2040 00:00:09 0x00000009 0x09 2041 00:00:10 0x0000000a 0x0a 2042 00:00:11 0x0000000b 0x0b 2043 00:00:12 0x0000000c 0x0c 2044 00:00:13 0x0000000d 0x0d 2045 00:00:14 0x0000000e 0x0e 2046 00:00:15 0x0000000f 0x0f 2047 00:00:16 0x00000010 0x10 2048 00:00:17 0x00000011 0x11 2049 00:00:18 0x00000012 0x12 2050 00:00:19 0x00000013 0x13 2051 00:00:20 0x00000014 0x14 2052 00:00:21 0x00000015 0x15 2053 00:00:22 0x00000016 0x16 2054 00:00:23 0x00000017 0x17 2055 00:00:24 0x00000018 0x18 2056 00:00:25 0x00000019 0x19 2057 00:00:26 0x0000001a 0x1a 2058 00:00:27 0x0000001b 0x1b 2059 00:00:28 0x0000001c 0x1c 2060 00:00:29 0x0000001d 0x1d 2061 00:00:30 0x0000001e 0x1e 2062 00:00:31 0x0000001f 0x1f 2063 00:00:32 0x00000020 0x20 2064 00:00:33 0x00000021 0x21 2065 00:00:34 0x00000022 0x22 2066 00:00:35 0x00000023 0x23 2067 00:00:36 0x00000024 0x24 2068 00:00:37 0x00000025 0x25 2069 00:00:38 0x00000026 0x26 2070 00:00:39 0x00000027 0x27 2071 00:00:40 0x00000028 0x28 2072 00:00:41 0x00000029 0x29 2073 00:00:42 0x0000002a 0x2a 2074 00:00:43 0x0000002b 0x2b 2075 00:00:44 0x0000002c 0x2c 2076 00:00:45 0x0000002d 0x2d 2077 00:00:46 0x0000002e 0x2e 2078 00:00:47 0x0000002f 0x2f 2079 00:00:48 0x00000030 0x30 2080 00:00:49 0x00000031 0x31 2081 00:00:50 0x00000032 0x32 2082 00:00:51 0x00000033 0x33 2083 00:00:52 0x00000034 0x34 2084 00:00:53 0x00000035 0x35 2085 00:00:54 0x00000036 0x36 2086 00:00:55 0x00000037 0x37 2087 00:00:56 0x00000038 0x38 2088 00:00:57 0x00000039 0x39 2089 00:00:58 0x0000003a 0x3a 2090 00:00:59 0x0000003b 0x3b 2091 00:01:00 0x0000003c 0x3c 2092 00:01:01 0x0000003d 0x3d 2093 00:01:02 0x0000003e 0x3e 2094 00:01:03 0x0000003f 0x3f 2095 00:01:04 0x00000040 0x40 2096 00:01:05 0x00000041 0x41 2097 00:01:06 0x00000042 0x42 2098 00:01:07 0x00000043 0x43 2099 00:01:08 0x00000044 0x44 2100 00:01:09 0x00000045 0x45 2101 00:01:10 0x00000046 0x46 2102 00:01:11 0x00000047 0x47 2103 00:01:12 0x00000048 0x48 2104 00:01:13 0x00000049 0x49 2105 00:01:14 0x0000004a 0x4a 2106 00:01:15 0x0000004b 0x4b 2107 00:01:16 0x0000004c 0x4c 2108 00:01:17 0x0000004d 0x4d 2109 00:01:18 0x0000004e 0x4e 2110 00:01:19 0x0000004f 0x4f 2111 00:01:20 0x00000050 0x50 2112 00:01:21 0x00000051 0x51 2113 00:01:22 0x00000052 0x52 2114 00:01:23 0x00000053 0x53 2115 00:01:24 0x00000054 0x54 2116 00:01:25 0x00000055 0x55 2117 00:01:26 0x00000056 0x56 2118 00:01:27 0x00000057 0x57 2119 00:01:28 0x00000058 0x58 2120 00:01:29 0x00000059 0x59 2121 00:01:30 0x0000005a 0x5a 2122 00:01:31 0x0000005b 0x5b 2123 00:01:32 0x0000005c 0x5c 2124 00:01:33 0x0000005d 0x5d 2125 00:01:34 0x0000005e 0x5e 2126 00:01:35 0x0000005f 0x5f 2127 00:01:36 0x00000060 0x60 2128 00:01:37 0x00000061 0x61 2129 00:01:38 0x00000062 0x62 2130 00:01:39 0x00000063 0x63 2131 00:01:40 0x00000064 0x64 2132 00:01:41 0x00000065 0x65 2133 00:01:42 0x00000066 0x66 2134 00:01:43 0x00000067 0x67 2135 00:01:44 0x00000068 0x68 2136 00:01:45 0x00000069 0x69 2137 00:01:46 0x0000006a 0x6a 2138 00:01:47 0x0000006b 0x6b 2139 00:01:48 0x0000006c 0x6c 2140 00:01:49 0x0000006d 0x6d 2141 00:01:50 0x0000006e 0x6e 2142 00:01:51 0x0000006f 0x6f 2143 00:01:52 0x00000070 0x70 2144 00:01:53 0x00000071 0x71 2145 00:01:54 0x00000072 0x72 2146 00:01:55 0x00000073 0x73 2147 00:01:56 0x00000074 0x74 2148 00:01:57 0x00000075 0x75 2149 00:01:58 0x00000076 0x76 2150 00:01:59 0x00000077 0x77 2151 00:02:00 0x00000078 0x78 2152 00:02:01 0x00000079 0x79 2153 00:02:02 0x0000007a 0x7a 2154 00:02:03 0x0000007b 0x7b 2155 00:02:04 0x0000007c 0x7c 2156 00:02:05 0x0000007d 0x7d 2157 00:02:06 0x0000007e 0x7e 2158 00:02:07 0x0000007f 0x7f 2159 00:02:08 0x00000080 0x80 2160 00:02:24 0x00000090 0x90 2161 00:02:40 0x000000a0 0xa0 2162 00:02:56 0x000000b0 0xb0 2163 00:03:12 0x000000c0 0xc0 2164 00:03:28 0x000000d0 0xd0 2165 00:03:44 0x000000e0 0xe0 2166 00:04:00 0x000000f0 0xf0 2167 00:04:16 0x00000100 0x81 2168 00:04:48 0x00000120 0x91 2169 00:05:20 0x00000140 0xa1 2170 00:05:52 0x00000160 0xb1 2171 00:06:24 0x00000180 0xc1 2172 00:06:56 0x000001a0 0xd1 2173 00:07:28 0x000001c0 0xe1 2174 00:08:00 0x000001e0 0xf1 2175 00:08:32 0x00000200 0x82 2176 00:09:36 0x00000240 0x92 2177 00:10:40 0x00000280 0xa2 2178 00:11:44 0x000002c0 0xb2 2179 00:12:48 0x00000300 0xc2 2180 00:13:52 0x00000340 0xd2 2181 00:14:56 0x00000380 0xe2 2182 00:16:00 0x000003c0 0xf2 2183 00:17:04 0x00000400 0x83 2184 00:19:12 0x00000480 0x93 2185 00:21:20 0x00000500 0xa3 2186 00:23:28 0x00000580 0xb3 2187 00:25:36 0x00000600 0xc3 2188 00:27:44 0x00000680 0xd3 2189 00:29:52 0x00000700 0xe3 2190 00:32:00 0x00000780 0xf3 2191 00:34:08 0x00000800 0x84 2192 00:38:24 0x00000900 0x94 2193 00:42:40 0x00000a00 0xa4 2194 00:46:56 0x00000b00 0xb4 2195 00:51:12 0x00000c00 0xc4 2196 00:55:28 0x00000d00 0xd4 2197 00:59:44 0x00000e00 0xe4 2198 01:04:00 0x00000f00 0xf4 2199 01:08:16 0x00001000 0x85 2200 01:16:48 0x00001200 0x95 2201 01:25:20 0x00001400 0xa5 2202 01:33:52 0x00001600 0xb5 2203 01:42:24 0x00001800 0xc5 2204 01:50:56 0x00001a00 0xd5 2205 01:59:28 0x00001c00 0xe5 2206 02:08:00 0x00001e00 0xf5 2207 02:16:32 0x00002000 0x86 2208 02:33:36 0x00002400 0x96 2209 02:50:40 0x00002800 0xa6 2210 03:07:44 0x00002c00 0xb6 2211 03:24:48 0x00003000 0xc6 2212 03:41:52 0x00003400 0xd6 2213 03:58:56 0x00003800 0xe6 2214 04:16:00 0x00003c00 0xf6 2215 04:33:04 0x00004000 0x87 2216 05:07:12 0x00004800 0x97 2217 05:41:20 0x00005000 0xa7 2218 06:15:28 0x00005800 0xb7 2219 06:49:36 0x00006000 0xc7 2220 07:23:44 0x00006800 0xd7 2221 07:57:52 0x00007000 0xe7 2222 08:32:00 0x00007800 0xf7 2223 09:06:08 0x00008000 0x88 2224 10:14:24 0x00009000 0x98 2225 11:22:40 0x0000a000 0xa8 2226 12:30:56 0x0000b000 0xb8 2227 13:39:12 0x0000c000 0xc8 2228 14:47:28 0x0000d000 0xd8 2229 15:55:44 0x0000e000 0xe8 2230 17:04:00 0x0000f000 0xf8 2231 18:12:16 0x00010000 0x89 2232 20:28:48 0x00012000 0x99 2233 22:45:20 0x00014000 0xa9 2234 1d 01:01:52 0x00016000 0xb9 2235 1d 03:18:24 0x00018000 0xc9 2236 1d 05:34:56 0x0001a000 0xd9 2237 1d 07:51:28 0x0001c000 0xe9 2238 1d 10:08:00 0x0001e000 0xf9 2239 1d 12:24:32 0x00020000 0x8a 2240 1d 16:57:36 0x00024000 0x9a 2241 1d 21:30:40 0x00028000 0xaa 2242 2d 02:03:44 0x0002c000 0xba 2243 2d 06:36:48 0x00030000 0xca 2244 2d 11:09:52 0x00034000 0xda 2245 2d 15:42:56 0x00038000 0xea 2246 2d 20:16:00 0x0003c000 0xfa 2247 3d 00:49:04 0x00040000 0x8b 2248 3d 09:55:12 0x00048000 0x9b 2249 3d 19:01:20 0x00050000 0xab 2250 4d 04:07:28 0x00058000 0xbb 2251 4d 13:13:36 0x00060000 0xcb 2252 4d 22:19:44 0x00068000 0xdb 2253 5d 07:25:52 0x00070000 0xeb 2254 5d 16:32:00 0x00078000 0xfb 2255 6d 01:38:08 0x00080000 0x8c 2256 6d 19:50:24 0x00090000 0x9c 2257 7d 14:02:40 0x000a0000 0xac 2258 8d 08:14:56 0x000b0000 0xbc 2259 9d 02:27:12 0x000c0000 0xcc 2260 9d 20:39:28 0x000d0000 0xdc 2261 10d 14:51:44 0x000e0000 0xec 2262 11d 09:04:00 0x000f0000 0xfc 2263 12d 03:16:16 0x00100000 0x8d 2264 13d 15:40:48 0x00120000 0x9d 2265 15d 04:05:20 0x00140000 0xad 2266 16d 16:29:52 0x00160000 0xbd 2267 18d 04:54:24 0x00180000 0xcd 2268 19d 17:18:56 0x001a0000 0xdd 2269 21d 05:43:28 0x001c0000 0xed 2270 22d 18:08:00 0x001e0000 0xfd 2271 24d 06:32:32 0x00200000 0x8e 2272 27d 07:21:36 0x00240000 0x9e 2273 30d 08:10:40 0x00280000 0xae 2274 33d 08:59:44 0x002c0000 0xbe 2275 36d 09:48:48 0x00300000 0xce 2276 39d 10:37:52 0x00340000 0xde 2277 42d 11:26:56 0x00380000 0xee 2278 45d 12:16:00 0x003c0000 0xfe 2279 48d 13:05:04 0x00400000 0x8f 2280 54d 14:43:12 0x00480000 0x9f 2281 60d 16:21:20 0x00500000 0xaf 2282 66d 17:59:28 0x00580000 0xbf 2283 72d 19:37:36 0x00600000 0xcf 2284 78d 21:15:44 0x00680000 0xdf 2285 84d 22:53:52 0x00700000 0xef 2286 91d 00:32:00 0x00780000 0xff (reserved) 2288 Figure 24 2290 Authors' Addresses 2292 Carsten Bormann 2293 Universitaet Bremen TZI 2294 Postfach 330440 2295 Bremen D-28359 2296 Germany 2298 Phone: +49-421-218-63921 2299 Email: cabo@tzi.org 2301 Klaus Hartke 2302 Universitaet Bremen TZI 2303 Postfach 330440 2304 Bremen D-28359 2305 Germany 2307 Phone: +49-421-218-63905 2308 Email: hartke@tzi.org