idnits 2.17.00 (12 Aug 2021) /tmp/idnits7625/draft-ietf-lpwan-coap-static-context-hc-12.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 both a reference to RFC 2119 and the recommended RFC 2119 boilerplate, even if it appears to use RFC 2119 keywords -- however, there's a paragraph with a matching beginning. Boilerplate error? RFC 2119 keyword, line 140: '... the CoAP header MAY be done in conjun...' RFC 2119 keyword, line 242: '...as found, then the packet MUST be sent...' RFC 2119 keyword, line 254: '...idirectional and MUST be elided during...' RFC 2119 keyword, line 268: '... The field SHOULD be elided if for i...' RFC 2119 keyword, line 271: '...any case, a rule MUST be defined to ca...' (20 more instances...) Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year == Line 1050 has weird spacing: '...tkn piv kid...' == Line 1067 has weird spacing: '... mid tkn...' -- The document date (December 10, 2019) is 893 days in the past. Is this intentional? Checking references for intended status: Proposed Standard ---------------------------------------------------------------------------- (See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) == Missing Reference: 'CON' is mentioned on line 264, but not defined == Missing Reference: 'NON' is mentioned on line 264, but not defined == Missing Reference: 'ACK' is mentioned on line 265, but not defined == Missing Reference: 'RST' is mentioned on line 265, but not defined -- Looks like a reference, but probably isn't: '69' on line 1091 -- Looks like a reference, but probably isn't: '132' on line 1091 == Outdated reference: draft-ietf-lpwan-ipv6-static-context-hc has been published as RFC 8724 Summary: 1 error (**), 0 flaws (~~), 8 warnings (==), 3 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 lpwan Working Group A. Minaburo 3 Internet-Draft Acklio 4 Intended status: Standards Track L. Toutain 5 Expires: June 12, 2020 Institut MINES TELECOM; IMT Atlantique 6 R. Andreasen 7 Universidad de Buenos Aires 8 December 10, 2019 10 LPWAN Static Context Header Compression (SCHC) for CoAP 11 draft-ietf-lpwan-coap-static-context-hc-12 13 Abstract 15 This draft defines the way SCHC header compression can be applied to 16 CoAP headers. The CoAP header structure differs from IPv6 and UDP 17 protocols since CoAP uses a flexible header with a variable number of 18 options, themselves of variable length. The CoAP protocol messages 19 format is asymmetric: the request messages have a header format 20 different from the one in the response messages. This document 21 explains how to use the SCHC compression mechanism for CoAP. 23 Status of This Memo 25 This Internet-Draft is submitted in full conformance with the 26 provisions of BCP 78 and BCP 79. 28 Internet-Drafts are working documents of the Internet Engineering 29 Task Force (IETF). Note that other groups may also distribute 30 working documents as Internet-Drafts. The list of current Internet- 31 Drafts is at https://datatracker.ietf.org/drafts/current/. 33 Internet-Drafts are draft documents valid for a maximum of six months 34 and may be updated, replaced, or obsoleted by other documents at any 35 time. It is inappropriate to use Internet-Drafts as reference 36 material or to cite them other than as "work in progress." 38 This Internet-Draft will expire on June 12, 2020. 40 Copyright Notice 42 Copyright (c) 2019 IETF Trust and the persons identified as the 43 document authors. All rights reserved. 45 This document is subject to BCP 78 and the IETF Trust's Legal 46 Provisions Relating to IETF Documents 47 (https://trustee.ietf.org/license-info) in effect on the date of 48 publication of this document. Please review these documents 49 carefully, as they describe your rights and restrictions with respect 50 to this document. Code Components extracted from this document must 51 include Simplified BSD License text as described in Section 4.e of 52 the Trust Legal Provisions and are provided without warranty as 53 described in the Simplified BSD License. 55 Table of Contents 57 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 2 58 1.1. Terminology . . . . . . . . . . . . . . . . . . . . . . . 3 59 2. SCHC Compression Process . . . . . . . . . . . . . . . . . . 3 60 3. CoAP Compression with SCHC . . . . . . . . . . . . . . . . . 4 61 4. Compression of CoAP header fields . . . . . . . . . . . . . . 6 62 4.1. CoAP version field . . . . . . . . . . . . . . . . . . . 6 63 4.2. CoAP type field . . . . . . . . . . . . . . . . . . . . . 6 64 4.3. CoAP code field . . . . . . . . . . . . . . . . . . . . . 6 65 4.4. CoAP Message ID field . . . . . . . . . . . . . . . . . . 7 66 4.5. CoAP Token fields . . . . . . . . . . . . . . . . . . . . 7 67 5. CoAP options . . . . . . . . . . . . . . . . . . . . . . . . 7 68 5.1. CoAP Content and Accept options. . . . . . . . . . . . . 7 69 5.2. CoAP option Max-Age, Uri-Host and Uri-Port fields . . . . 8 70 5.3. CoAP option Uri-Path and Uri-Query fields . . . . . . . . 8 71 5.3.1. Variable length Uri-Path and Uri-Query . . . . . . . 9 72 5.3.2. Variable number of path or query elements . . . . . . 9 73 5.4. CoAP option Size1, Size2, Proxy-URI and Proxy-Scheme 74 fields . . . . . . . . . . . . . . . . . . . . . . . . . 9 75 5.5. CoAP option ETag, If-Match, If-None-Match, Location-Path 76 and Location-Query fields . . . . . . . . . . . . . . . . 10 77 6. Other RFCs . . . . . . . . . . . . . . . . . . . . . . . . . 10 78 6.1. Block . . . . . . . . . . . . . . . . . . . . . . . . . . 10 79 6.2. Observe . . . . . . . . . . . . . . . . . . . . . . . . . 10 80 6.3. No-Response . . . . . . . . . . . . . . . . . . . . . . . 10 81 6.4. OSCORE . . . . . . . . . . . . . . . . . . . . . . . . . 11 82 7. Examples of CoAP header compression . . . . . . . . . . . . . 12 83 7.1. Mandatory header with CON message . . . . . . . . . . . . 12 84 7.2. OSCORE Compression . . . . . . . . . . . . . . . . . . . 13 85 7.3. Example OSCORE Compression . . . . . . . . . . . . . . . 16 86 8. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 26 87 9. Security considerations . . . . . . . . . . . . . . . . . . . 26 88 10. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 26 89 11. Normative References . . . . . . . . . . . . . . . . . . . . 26 90 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 27 92 1. Introduction 94 CoAP [rfc7252] is an implementation of the REST architecture for 95 constrained devices. Although CoAP was designed for constrained 96 devices, the size of a CoAP header still is too large for the 97 constraints of Low Power Wide Area Networks (LPWAN) and some 98 compression is needed to reduce the header size. 100 [I-D.ietf-lpwan-ipv6-static-context-hc] defines a header compression 101 mechanism for LPWAN network based on a static context. The context 102 is said static since the field description composing the Rules are 103 not learned during the packet exchanges but are previously defined. 104 The context(s) is(are) known by both ends before transmission. 106 A context is composed of a set of rules that are referenced by Rule 107 IDs (identifiers). A rule contains an ordered list of the fields 108 descriptions containing a field ID (FID), its length (FL) and its 109 position (FP), a direction indicator (DI) (upstream, downstream and 110 bidirectional) and some associated Target Values (TV). Target Value 111 indicates the value that can be expected. TV can also be a list of 112 values. A Matching Operator (MO) is associated to each header field 113 description. The rule is selected if all the MOs fit the TVs for all 114 fields of the incoming packet. In that case, a Compression/ 115 Decompression Action (CDA) associated to each field defines how the 116 compressed and the decompressed values are computed out of each 117 other, for each of the header fields. Compression mainly results in 118 one of 4 actions: send the field value, send nothing, send some least 119 significant bits of the field or send an index. After applying the 120 compression there may be some bits to be sent, these values are 121 called Compression Residues and are transmitted after the Rule ID in 122 the compressed messages. 124 The compression rules define a generic way to compress and decompress 125 the fields. If the device is modified, for example, to introduce new 126 functionalities or new CoAP options, the rules must be updated to 127 reflect the evolution. 129 1.1. Terminology 131 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 132 "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and 133 "OPTIONAL" in this document are to be interpreted as described in BCP 134 14 [rfc2119][rfc8174] when, and only when, they appear in all 135 capitals, as shown here. 137 2. SCHC Compression Process 139 The SCHC Compression rules can be applied to CoAP flows. SCHC 140 Compression of the CoAP header MAY be done in conjunction with the 141 lower layers (IPv6/UDP) or independently. The SCHC adaptation layers 142 as described in [I-D.ietf-lpwan-ipv6-static-context-hc] may be used 143 as shown in Figure 1. 145 ^ +------------+ ^ +------------+ ^ +------------+ 146 | | CoAP | | | CoAP | inner | | CoAP | 147 | +------------+ v +------------+ x | OSCORE | 148 | | UDP | | DTLS | outer | +------------+ 149 | +------------+ +------------+ | | UDP | 150 | | IPv6 | | UDP | | +------------+ 151 v +------------+ +------------+ | | IPv6 | 152 | IPv6 | v +------------+ 153 +------------+ 155 Figure 1: rule scope for CoAP 157 Figure 1 shows some examples for CoAP architecture and the SCHC 158 rule's scope. 160 In the first example, a rule compresses the complete header stack 161 from IPv6 to CoAP. In this case, SCHC C/D (Static Context Header 162 Compression Compressor/Decompressor) is performed at the device and 163 at the LPWAN boundary. 165 In the second example, an end-to-end encryption mechanisms is used 166 between the device and the application. The SCHC compression is 167 applied in the CoAP layer compressing the CoAP header independently 168 of the other layers. The rule ID and the compression residue are 169 encrypted using a mechanism such as DTLS. Only the other end can 170 decipher the information. Layers below may also be compressed using 171 other SCHC rules (this is out of the scope of this document) as 172 defined in the SCHC [I-D.ietf-lpwan-ipv6-static-context-hc] document. 174 In the third example, OSCORE [rfc8613] is used. In this case, two 175 rulesets are used to compress the CoAP message. A first ruleset 176 focused on the inner header and is applied end to end by both ends. 177 A second ruleset compresses the outer header and the layers below and 178 is done between the device and the LPWAN boundary. 180 3. CoAP Compression with SCHC 182 CoAP differs from IPv6 and UDP protocols on the following aspects: 184 o IPv6 and UDP are symmetrical protocols. The same fields are found 185 in the request and in the response, with the value of some fields 186 being swapped on the return path (e.g. source and destination 187 fields). A CoAP request is intrinsically different from a 188 response. For example, the URI-path option is mandatory in the 189 request and is not found in the response, a request may contain an 190 Accept option and the response a Content option. 192 [I-D.ietf-lpwan-ipv6-static-context-hc] defines the use of a 193 message direction (DI) in the Field Description, which allows a 194 single Rule to process message headers differently depending of 195 the direction. 197 o Even when a field is "symmetric" (i.e. found in both directions) 198 the values carried in each direction are different. Combined with 199 a matching list in the TV, this allows reducing the range of 200 expected values in a particular direction and therefore reduce the 201 size of the compression residue. For instance, if a client sends 202 only CON request, the type can be elided by compression and the 203 answer may use one single bit to carry either the ACK or RST type. 204 The same behavior can be applied to the CoAP Code field 0.0X code 205 Format is found in the request and Y.ZZ code format in the answer. 206 The direction allows splitting in two parts the possible values 207 for each direction in the same Rule. 209 o In IPv6 and UDP, header fields have a fixed size and it is not 210 sent. In CoAP, some fields in the header have a varying size, for 211 example the Token size may vary from 0 to 8 bytes, the length is 212 given by a field in the header. More systematically, the CoAP 213 options are described using the Type-Length-Value. 215 [I-D.ietf-lpwan-ipv6-static-context-hc] offers the possibility to 216 define a function for the Field Length in the Field Description. 218 o In CoAP headers, a field can appear several times. This is 219 typical for elements of a URI (path or queries). The SCHC 220 specification [I-D.ietf-lpwan-ipv6-static-context-hc] allows a 221 Field ID to appears several times in the rule, and uses the Field 222 Position (FP) to identify the correct instance, and thereby 223 removing the ambiguity of the matching operation. 225 o Field sizes defined in the CoAP protocol can be too large 226 regarding LPWAN traffic constraints. This is particularly true 227 for the Message ID field and the Token field. The MSB MO can be 228 applied to reduce the information carried on LPWANs. 230 o CoAP also obeys the client/server paradigm and the compression 231 ratio can be different if the request is issued from an LPWAN 232 device or from a non LPWAN device. For instance, a Device (Dev) 233 aware of LPWAN constraints can generate a 1-byte token, but a 234 regular CoAP client will certainly send a larger token to the Dev. 235 The SCHC compression-decompression process never modifies the 236 Values it only reduces their sizes. Nevertheless, a proxy placed 237 before the compressor may change some field values to allow SCHC 238 achieving a better compression ratio, while maintaining the 239 necessary context for interoperability with existing CoAP 240 implementations. 242 o If no valid Rule was found, then the packet MUST be sent 243 uncompressed using the RuleID dedicated to this purpose and the 244 Compression Residue is the complete header of the packet. See 245 section 6 of [I-D.ietf-lpwan-ipv6-static-context-hc]. 247 4. Compression of CoAP header fields 249 This section discusses the compression of the different CoAP header 250 fields. 252 4.1. CoAP version field 254 CoAP version is bidirectional and MUST be elided during the SCHC 255 compression, since it always contains the same value. In the future, 256 if new versions of CoAP are defined, new rules will be needed to 257 avoid ambiguities between versions. 259 4.2. CoAP type field 261 CoAP Protocol [rfc7252] defines 4 types of messages: CON, NON, ACK 262 and RST. ACK and RST are a response to the CON and NON. If the 263 device plays a specific client or server role, a rule can take 264 advantage of these properties with the mapping list: [CON, NON] for 265 one direction and [ACK, RST] for the other direction and so, the 266 compression residue is reduced to 1 bit. 268 The field SHOULD be elided if for instance a client is sending only 269 NON or only CON messages. 271 In any case, a rule MUST be defined to carry RST to a client. 273 4.3. CoAP code field 275 The compression of the CoAP code field follows the same principle as 276 that of the CoAP type field. If the device plays a specific role, 277 the set of code values can be split in two parts, the request codes 278 with the 0 class and the response values. 280 If the device only implements a CoAP client, the request code can be 281 reduced to the set of requests the client is able to process. 283 All the response codes MUST be compressed with a SCHC rule. 285 4.4. CoAP Message ID field 287 The Message ID field is bidirectional and is used to manage 288 acknowledgments. The server memorizes the value for an 289 EXCHANGE_LIFETIME period (by default 247 seconds) for CON messages 290 and a NON_LIFETIME period (by default 145 seconds) for NON messages. 291 During that period, a server receiving the same Message ID value will 292 process the message as a retransmission. After this period, it will 293 be processed as a new message. 295 In case where the Device is a client, the size of the Message ID 296 field may be too large regarding the number of messages sent. The 297 client SHOULD use only small Message ID values, for instance 4 bit 298 long. Therefore, an MSB can be used to limit the size of the 299 compression residue. 301 In case where the Device is a server, the client may be located 302 outside of the LPWAN area and it views the Device as a regular device 303 connected to the Internet. The client will generate Message ID using 304 the 16 bits space offered by this field. A CoAP proxy can be set 305 before the SCHC C/D to reduce the value of the Message ID, to allow 306 its compression with the MSB matching operator and LSB CDA. 308 4.5. CoAP Token fields 310 Token is defined through two CoAP fields, Token Length in the 311 mandatory header and Token Value directly following the mandatory 312 CoAP header. 314 Token Length is processed as any protocol field. If the value 315 remains the same during all the transaction, the size can be stored 316 in the context and elided during the transmission. Otherwise, it 317 will have to be sent as a compression residue. 319 Token Value size cannot be defined directly in the rule in the Field 320 Length (FL). Instead, a specific function designated as "TKL" MUST 321 be used and length does not have to be sent with the residue. During 322 the decompression, this function returns the value contained in the 323 Token Length field. 325 5. CoAP options 327 5.1. CoAP Content and Accept options. 329 These fields are both unidirectional and MUST NOT be set to 330 bidirectional in a rule entry. 332 If a single value is expected by the client, it can be stored in the 333 TV and elided during the transmission. Otherwise, if several 334 possible values are expected by the client, a matching-list SHOULD be 335 used to limit the size of the residue. Otherwise, the value has to 336 be sent as a residue (fixed or variable length). 338 5.2. CoAP option Max-Age, Uri-Host and Uri-Port fields 340 These fields are unidirectional and MUST NOT be set to bidirectional 341 in a rule entry. They are used only by the server to inform of the 342 caching duration and is never found in client requests. 344 If the duration is known by both ends, the value can be elided on the 345 LPWAN. 347 A matching list can be used if some well-known values are defined. 349 Otherwise these options SHOULD be sent as a residue (fixed or 350 variable length). 352 5.3. CoAP option Uri-Path and Uri-Query fields 354 These fields are unidirectional and MUST NOT be set to bidirectional 355 in a rule entry. They are used only by the client to access a 356 specific resource and are never found in server responses. 358 Uri-Path and Uri-Query elements are a repeatable options, the Field 359 Position (FP) gives the position in the path. 361 A Mapping list can be used to reduce the size of variable Paths or 362 Queries. In that case, to optimize the compression, several elements 363 can be regrouped into a single entry. Numbering of elements do not 364 change, MO comparison is set with the first element of the matching. 366 +-------------+--+--+--+--------+---------+-------------+ 367 | Field |FL|FP|DI| Target | Match | CDA | 368 | | | | | Value | Opera. | | 369 +-------------+--+--+--+--------+---------+-------------+ 370 |URI-Path | | 1|up|["/a/b",|equal |not-sent | 371 | | | | |"/c/d"] | | | 372 |URI-Path | | 3|up| |ignore |value-sent | 373 +-------------+--+--+--+--------+---------+-------------+ 375 Figure 2: complex path example 377 In Figure 2 a single bit residue can be used to code one of the 2 378 paths. If regrouping were not allowed, a 2 bits residue would be 379 needed. 381 5.3.1. Variable length Uri-Path and Uri-Query 383 When the length is not known at the rule creation, the Field Length 384 SHOULD be set to variable, and the unit is set to bytes. 386 The MSB MO can be applied to a Uri-Path or Uri-Query element. Since 387 MSB value is given in bit, the size MUST always be a multiple of 8 388 bits. 390 The length sent at the beginning of a variable length residue 391 indicates the size of the LSB in bytes. 393 For instance for a CORECONF path /c/X6?k="eth0" the rule can be set 394 to: 396 +-------------+---+--+--+--------+---------+-------------+ 397 | Field |FL |FP|DI| Target | Match | CDA | 398 | | | | | Value | Opera. | | 399 +-------------+---+--+--+--------+---------+-------------+ 400 |URI-Path | 8| 1|up|"c" |equal |not-sent | 401 |URI-Path |var| 2|up| |ignore |value-sent | 402 |URI-Query |var| 1|up|"k=" |MSB(16) |LSB | 403 +-------------+---+--+--+--------+---------+-------------+ 405 Figure 3: CORECONF URI compression 407 Figure 3 shows the parsing and the compression of the URI, where c is 408 not sent. The second element is sent with the length (i.e. 0x2 X 6) 409 followed by the query option (i.e. 0x05 "eth0"). 411 5.3.2. Variable number of path or query elements 413 The number of Uri-path or Uri-Query elements in a rule is fixed at 414 the rule creation time. If the number varies, several rules SHOULD 415 be created to cover all the possibilities. Another possibility is to 416 define the length of Uri-Path to variable and send a compression 417 residue with a length of 0 to indicate that this Uri-Path is empty. 418 This adds 4 bits to the compression residue. 420 5.4. CoAP option Size1, Size2, Proxy-URI and Proxy-Scheme fields 422 These fields are unidirectional and MUST NOT be set to bidirectional 423 in a rule entry. They are used only by the client to access a 424 specific resource and are never found in server response. 426 If the field value has to be sent, TV is not set, MO is set to 427 "ignore" and CDA is set to "value-sent". A mapping MAY also be used. 429 Otherwise, the TV is set to the value, MO is set to "equal" and CDA 430 is set to "not-sent". 432 5.5. CoAP option ETag, If-Match, If-None-Match, Location-Path and 433 Location-Query fields 435 These fields are unidirectional. 437 These fields values cannot be stored in a rule entry. They MUST 438 always be sent with the compression residues. 440 6. Other RFCs 442 6.1. Block 444 Block [rfc7959] allows a fragmentation at the CoAP level. SCHC also 445 includes a fragmentation protocol. They are compatible. If a block 446 option is used, its content MUST be sent as a compression residue. 448 6.2. Observe 450 The [rfc7641] defines the Observe option. The TV is not set, MO is 451 set to "ignore" and the CDA is set to "value-sent". SCHC does not 452 limit the maximum size for this option (3 bytes). To reduce the 453 transmission size, either the device implementation MAY limit the 454 delta between two consecutive values, or a proxy can modify the 455 increment. 457 Since an RST message may be sent to inform a server that the client 458 does not require Observe response, a rule MUST allow the transmission 459 of this message. 461 6.3. No-Response 463 The [rfc7967] defines a No-Response option limiting the responses 464 made by a server to a request. If the value is known by both ends, 465 then TV is set to this value, MO is set to "equal" and CDA is set to 466 "not-sent". 468 Otherwise, if the value is changing over time, TV is not set, MO is 469 set to "ignore" and CDA to "value-sent". A matching list can also be 470 used to reduce the size. 472 6.4. OSCORE 474 OSCORE [rfc8613] defines end-to-end protection for CoAP messages. 475 This section describes how SCHC rules can be applied to compress 476 OSCORE-protected messages. 478 0 1 2 3 4 5 6 7 <--------- n bytes -------------> 479 +-+-+-+-+-+-+-+-+--------------------------------- 480 |0 0 0|h|k| n | Partial IV (if any) ... 481 +-+-+-+-+-+-+-+-+--------------------------------- 482 | | | 483 |<-- CoAP -->|<------ CoAP OSCORE_piv ------> | 484 OSCORE_flags 486 <- 1 byte -> <------ s bytes -----> 487 +------------+----------------------+-----------------------+ 488 | s (if any) | kid context (if any) | kid (if any) ... | 489 +------------+----------------------+-----------------------+ 490 | | | 491 | <------ CoAP OSCORE_kidctxt ----->|<-- CoAP OSCORE_kid -->| 493 Figure 4: OSCORE Option 495 The encoding of the OSCORE Option Value defined in Section 6.1 of 496 [rfc8613] is repeated in Figure 4. 498 The first byte is used for flags that specify the contents of the 499 OSCORE option. The 3 most significant bits of this byte are reserved 500 and always set to 0. Bit h, when set, indicates the presence of the 501 kid context field in the option. Bit k, when set, indicates the 502 presence of a kid field. The 3 least significant bits n indicate the 503 length of the piv (Partial Initialization Vector) field in bytes. 504 When n = 0, no piv is present. 506 The flag byte is followed by the piv field, kid context field and kid 507 field in this order and if present; the length of the kid context 508 field is encoded in the first byte denoting by s the length of the 509 kid context in bytes. 511 This draft recommends to implement a parser that is able to identify 512 the OSCORE Option and the fields it contains. 514 Conceptually, it discerns up to 4 distinct pieces of information 515 within the OSCORE option: the flag bits, the piv, the kid context, 516 and the kid. It is thus recommended that the parser split the OSCORE 517 option into the 4 subsequent fields: 519 o CoAP OSCORE_flags, 521 o CoAP OSCORE_piv, 523 o CoAP OSCORE_kidctxt, 525 o CoAP OSCORE_kid. 527 These fields are shown superimposed on the OSCORE Option format in 528 Figure 4, the CoAP OSCORE_kidctxt field including the size bits s. 529 Their size SHOULD be reduced using SCHC compression. 531 7. Examples of CoAP header compression 533 7.1. Mandatory header with CON message 535 In this first scenario, the LPWAN compressor at the Network Gateway 536 side receives from an Internet client a POST message, which is 537 immediately acknowledged by the Device. For this simple scenario, 538 the rules are described Figure 5. 540 Rule ID 1 541 +-------------+--+--+--+------+---------+-------------++------------+ 542 | Field |FL|FP|DI|Target| Match | CDA || Sent | 543 | | | | |Value | Opera. | || [bits] | 544 +-------------+--+--+--+------+---------+-------------++------------+ 545 |CoAP version | | |bi| 01 |equal |not-sent || | 546 |CoAP Type | | |dw| CON |equal |not-sent || | 547 |CoAP Type | | |up|[ACK, | | || | 548 | | | | | RST] |match-map|matching-sent|| T | 549 |CoAP TKL | | |bi| 0 |equal |not-sent || | 550 |CoAP Code | | |bi|[0.00,| | || | 551 | | | | | ... | | || | 552 | | | | | 5.05]|match-map|matching-sent|| CC CCC | 553 |CoAP MID | | |bi| 0000 |MSB(7 ) |LSB || M-ID| 554 |CoAP Uri-Path| | |dw| path |equal 1 |not-sent || | 555 +-------------+--+--+--+------+---------+-------------++------------+ 557 Figure 5: CoAP Context to compress header without token 559 The version and Token Length fields are elided. The 26 method and 560 response codes defined in [rfc7252] has been shrunk to 5 bits using a 561 matching list. Uri-Path contains a single element indicated in the 562 matching operator. 564 SCHC Compression reduces the header sending only the Type, a mapped 565 code and the least significant bits of Message ID (9 bits in the 566 example above). 568 Note that a request sent by a client located in an Application Server 569 to a server located in the device, may not be compressed through this 570 rule since the MID will not start with 7 bits equal to 0. A CoAP 571 proxy, before the core SCHC C/D can rewrite the message ID to a value 572 matched by the rule. 574 7.2. OSCORE Compression 576 OSCORE aims to solve the problem of end-to-end encryption for CoAP 577 messages. The goal, therefore, is to hide as much of the message as 578 possible while still enabling proxy operation. 580 Conceptually this is achieved by splitting the CoAP message into an 581 Inner Plaintext and Outer OSCORE Message. The Inner Plaintext 582 contains sensitive information which is not necessary for proxy 583 operation. This, in turn, is the part of the message which can be 584 encrypted until it reaches its end destination. The Outer Message 585 acts as a shell matching the format of a regular CoAP message, and 586 includes all Options and information needed for proxy operation and 587 caching. This decomposition is illustrated in Figure 6. 589 CoAP options are sorted into one of 3 classes, each granted a 590 specific type of protection by the protocol: 592 o Class E: Encrypted options moved to the Inner Plaintext, 594 o Class I: Integrity-protected options included in the AAD for the 595 encryption of the Plaintext but otherwise left untouched in the 596 Outer Message, 598 o Class U: Unprotected options left untouched in the Outer Message. 600 Additionally, the OSCORE Option is added as an Outer option, 601 signalling that the message is OSCORE protected. This option carries 602 the information necessary to retrieve the Security Context with which 603 the message was encrypted so that it may be correctly decrypted at 604 the other end-point. 606 Original CoAP Message 607 +-+-+---+-------+---------------+ 608 |v|t|tkl| code | Msg Id. | 609 +-+-+---+-------+---------------+....+ 610 | Token | 611 +-------------------------------.....+ 612 | Options (IEU) | 613 . . 614 . . 615 +------+-------------------+ 616 | 0xFF | 617 +------+------------------------+ 618 | | 619 | Payload | 620 | | 621 +-------------------------------+ 622 / \ 623 / \ 624 / \ 625 / \ 626 Outer Header v v Plaintext 627 +-+-+---+--------+---------------+ +-------+ 628 |v|t|tkl|new code| Msg Id. | | code | 629 +-+-+---+--------+---------------+....+ +-------+-----......+ 630 | Token | | Options (E) | 631 +--------------------------------.....+ +-------+------.....+ 632 | Options (IU) | | OxFF | 633 . . +-------+-----------+ 634 . OSCORE Option . | | 635 +------+-------------------+ | Payload | 636 | 0xFF | | | 637 +------+ +-------------------+ 639 Figure 6: A CoAP message is split into an OSCORE outer and plaintext 641 Figure 6 shows the message format for the OSCORE Message and 642 Plaintext. 644 In the Outer Header, the original message code is hidden and replaced 645 by a default dummy value. As seen in sections 4.1.3.5 and 4.2 of the 646 [rfc8613], the message code is replaced by POST for requests and 647 Changed for responses when Observe is not used. If Observe is used, 648 the message code is replaced by FETCH for requests and Content for 649 responses. 651 The original message code is put into the first byte of the 652 Plaintext. Following the message code, the class E options comes and 653 if present the original message Payload is preceded by its payload 654 marker. 656 The Plaintext is now encrypted by an AEAD algorithm which integrity 657 protects Security Context parameters and eventually any class I 658 options from the Outer Header. Currently no CoAP options are marked 659 class I. The resulting Ciphertext becomes the new Payload of the 660 OSCORE message, as illustrated in Figure 7. 662 This Ciphertext is, as defined in RFC 5116, the concatenation of the 663 encrypted Plaintext and its authentication tag. Note that Inner 664 Compression only affects the Plaintext before encryption, thus we can 665 only aim to reduce this first, variable length component of the 666 Ciphertext. The authentication tag is fixed in length and considered 667 part of the cost of protection. 669 Outer Header 670 +-+-+---+--------+---------------+ 671 |v|t|tkl|new code| Msg Id. | 672 +-+-+---+--------+---------------+....+ 673 | Token | 674 +--------------------------------.....+ 675 | Options (IU) | 676 . . 677 . OSCORE Option . 678 +------+-------------------+ 679 | 0xFF | 680 +------+---------------------------+ 681 | | 682 | Ciphertext: Encrypted Inner | 683 | Header and Payload | 684 | + Authentication Tag | 685 | | 686 +----------------------------------+ 688 Figure 7: OSCORE message 690 The SCHC Compression scheme consists of compressing both the 691 Plaintext before encryption and the resulting OSCORE message after 692 encryption, see Figure 8. 694 This translates into a segmented process where SCHC compression is 695 applied independently in 2 stages, each with its corresponding set of 696 rules, with the Inner SCHC Rules and the Outer SCHC Rules. This way 697 compression is applied to all fields of the original CoAP message. 699 Note that since the Inner part of the message can only be decrypted 700 by the corresponding end-point, this end-point will also have to 701 implement Inner SCHC Compression/Decompression. 703 Outer Message OSCORE Plaintext 704 +-+-+---+--------+---------------+ +-------+ 705 |v|t|tkl|new code| Msg Id. | | code | 706 +-+-+---+--------+---------------+....+ +-------+-----......+ 707 | Token | | Options (E) | 708 +--------------------------------.....+ +-------+------.....+ 709 | Options (IU) | | OxFF | 710 . . +-------+-----------+ 711 . OSCORE Option . | | 712 +------+-------------------+ | Payload | 713 | 0xFF | | | 714 +------+------------+ +-------------------+ 715 | Ciphertext |<---------\ | 716 | | | v 717 +-------------------+ | +-----------------+ 718 | | | Inner SCHC | 719 v | | Compression | 720 +-----------------+ | +-----------------+ 721 | Outer SCHC | | | 722 | Compression | | v 723 +-----------------+ | +-------+ 724 | | |Rule ID| 725 v | +-------+--+ 726 +--------+ +------------+ | Residue | 727 |Rule ID'| | Encryption | <--- +----------+--------+ 728 +--------+--+ +------------+ | | 729 | Residue' | | Payload | 730 +-----------+-------+ | | 731 | Ciphertext | +-------------------+ 732 | | 733 +-------------------+ 735 Figure 8: OSCORE Compression Diagram 737 7.3. Example OSCORE Compression 739 An example is given with a GET Request and its consequent CONTENT 740 Response from a device-based CoAP client to a cloud-based CoAP 741 server. A possible set of rules for the Inner and Outer SCHC 742 Compression is shown. A dump of the results and a contrast between 743 SCHC + OSCORE performance with SCHC + COAP performance is also 744 listed. This gives an approximation to the cost of security with 745 SCHC-OSCORE. 747 Our first example CoAP message is the GET Request in Figure 9 749 Original message: 750 ================= 751 0x4101000182bb74656d7065726174757265 753 Header: 754 0x4101 755 01 Ver 756 00 CON 757 0001 tkl 758 00000001 Request Code 1 "GET" 760 0x0001 = mid 761 0x82 = token 763 Options: 764 0xbb74656d7065726174757265 765 Option 11: URI_PATH 766 Value = temperature 768 Original msg length: 17 bytes. 770 Figure 9: CoAP GET Request 772 Its corresponding response is the CONTENT Response in Figure 10. 774 Original message: 775 ================= 776 0x6145000182ff32332043 778 Header: 779 0x6145 780 01 Ver 781 10 ACK 782 0001 tkl 783 01000101 Successful Response Code 69 "2.05 Content" 785 0x0001 = mid 786 0x82 = token 788 0xFF Payload marker 789 Payload: 790 0x32332043 792 Original msg length: 10 794 Figure 10: CoAP CONTENT Response 796 The SCHC Rules for the Inner Compression include all fields that are 797 already present in a regular CoAP message, what is important is their 798 order and the definition of only those CoAP fields are into 799 Plaintext, Figure 11. 801 Rule ID 0 802 +---------------+--+--+-----------+-----------+-----------++------+ 803 | Field |FP|DI| Target | MO | CDA || Sent | 804 | | | | Value | | ||[bits]| 805 +---------------+--+--+-----------+-----------+-----------++------+ 806 |CoAP Code | |up| 1 | equal |not-sent || | 807 |CoAP Code | |dw|[69,132] | match-map |match-sent || c | 808 |CoAP Uri-Path | |up|temperature| equal |not-sent || | 809 |COAP Option-End| |dw| 0xFF | equal |not-sent || | 810 +---------------+--+--+-----------+-----------+-----------++------+ 812 Figure 11: Inner SCHC Rules 814 Figure 12 shows the Plaintext obtained for our example GET Request 815 and follows the process of Inner Compression and Encryption until we 816 end up with the Payload to be added in the outer OSCORE Message. 818 In this case the original message has no payload and its resulting 819 Plaintext can be compressed up to only 1 byte (size of the Rule ID). 820 The AEAD algorithm preserves this length in its first output, but 821 also yields a fixed-size tag which cannot be compressed and has to be 822 included in the OSCORE message. This translates into an overhead in 823 total message length, which limits the amount of compression that can 824 be achieved and plays into the cost of adding security to the 825 exchange. 827 ________________________________________________________ 828 | | 829 | OSCORE Plaintext | 830 | | 831 | 0x01bb74656d7065726174757265 (13 bytes) | 832 | | 833 | 0x01 Request Code GET | 834 | | 835 | bb74656d7065726174757265 Option 11: URI_PATH | 836 | Value = temperature | 837 |________________________________________________________| 839 | 840 | 841 | Inner SCHC Compression 842 | 843 v 844 _________________________________ 845 | | 846 | Compressed Plaintext | 847 | | 848 | 0x00 | 849 | | 850 | Rule ID = 0x00 (1 byte) | 851 | (No residue) | 852 |_________________________________| 854 | 855 | AEAD Encryption 856 | (piv = 0x04) 857 v 858 _________________________________________________ 859 | | 860 | encrypted_plaintext = 0xa2 (1 byte) | 861 | tag = 0xc54fe1b434297b62 (8 bytes) | 862 | | 863 | ciphertext = 0xa2c54fe1b434297b62 (9 bytes) | 864 |_________________________________________________| 866 Figure 12: Plaintext compression and encryption for GET Request 868 In Figure 13 we repeat the process for the example CONTENT Response. 869 In this case the misalignment produced by the compression residue (1 870 bit) makes it so that 7 bits of padding have to be applied after the 871 payload, resulting in a compressed Plaintext that is the same size as 872 before compression. This misalignment also causes the hexcode from 873 the payload to differ from the original, even though it has not been 874 compressed. On top of this, the overhead from the tag bytes is 875 incurred as before. 877 ________________________________________________________ 878 | | 879 | OSCORE Plaintext | 880 | | 881 | 0x45ff32332043 (6 bytes) | 882 | | 883 | 0x45 Successful Response Code 69 "2.05 Content" | 884 | | 885 | ff Payload marker | 886 | | 887 | 32332043 Payload | 888 |________________________________________________________| 890 | 891 | 892 | Inner SCHC Compression 893 | 894 v 895 __________________________________________ 896 | | 897 | Compressed Plaintext | 898 | | 899 | 0x001919902180 (6 bytes) | 900 | | 901 | 00 Rule ID | 902 | | 903 | 0b0 (1 bit match-map residue) | 904 | 0x32332043 >> 1 (shifted payload) | 905 | 0b0000000 Padding | 906 |__________________________________________| 908 | 909 | AEAD Encryption 910 | (piv = 0x04) 911 v 912 _________________________________________________________ 913 | | 914 | encrypted_plaintext = 0x10c6d7c26cc1 (6 bytes) | 915 | tag = 0xe9aef3f2461e0c29 (8 bytes) | 916 | | 917 | ciphertext = 0x10c6d7c26cc1e9aef3f2461e0c29 (14 bytes) | 918 |_________________________________________________________| 920 Figure 13: Plaintext compression and encryption for CONTENT Response 921 The Outer SCHC Rules (Figure 16) MUST process the OSCORE Options 922 fields. In Figure 14 and Figure 15 we show a dump of the OSCORE 923 Messages generated from our example messages once they have been 924 provided with the Inner Compressed Ciphertext in the payload. These 925 are the messages that have to be compressed by the Outer SCHC 926 Compression. 928 Protected message: 929 ================== 930 0x4102000182d7080904636c69656e74ffa2c54fe1b434297b62 931 (25 bytes) 933 Header: 934 0x4102 935 01 Ver 936 00 CON 937 0001 tkl 938 00000010 Request Code 2 "POST" 940 0x0001 = mid 941 0x82 = token 943 Options: 944 0xd8080904636c69656e74 (10 bytes) 945 Option 21: OBJECT_SECURITY 946 Value = 0x0904636c69656e74 947 09 = 000 0 1 001 Flag byte 948 h k n 949 04 piv 950 636c69656e74 kid 952 0xFF Payload marker 953 Payload: 954 0xa2c54fe1b434297b62 (9 bytes) 956 Figure 14: Protected and Inner SCHC Compressed GET Request 958 Protected message: 959 ================== 960 0x6144000182d008ff10c6d7c26cc1e9aef3f2461e0c29 961 (22 bytes) 963 Header: 964 0x6144 965 01 Ver 966 10 ACK 967 0001 tkl 968 01000100 Successful Response Code 68 "2.04 Changed" 970 0x0001 = mid 971 0x82 = token 973 Options: 974 0xd008 (2 bytes) 975 Option 21: OBJECT_SECURITY 976 Value = b'' 978 0xFF Payload marker 979 Payload: 980 0x10c6d7c26cc1e9aef3f2461e0c29 (14 bytes) 982 Figure 15: Protected and Inner SCHC Compressed CONTENT Response 984 For the flag bits, a number of compression methods has been shown to 985 be useful depending on the application. The simplest alternative is 986 to provide a fixed value for the flags, combining MO equal and CDA 987 not- sent. This saves most bits but could prevent flexibility. 988 Otherwise, match-mapping could be used to choose from an interested 989 number of configurations to the exchange. Otherwise, MSB could be 990 used to mask off the 3 hard-coded most significant bits. 992 Note that fixing a flag bit will limit the choice of CoAP Options 993 that can be used in the exchange, since their values are dependent on 994 certain options. 996 The piv field lends itself to having a number of bits masked off with 997 MO MSB and CDA LSB. This could be useful in applications where the 998 message frequency is low such as that found in LPWAN technologies. 999 Note that compressing the sequence numbers effectively reduces the 1000 maximum amount of sequence numbers that can be used in an exchange. 1001 Once this amount is exceeded, the SCHC Context would need to be re- 1002 established. 1004 The size s included in the kid context field MAY be masked off with 1005 CDA MSB. The rest of the field could have additional bits masked 1006 off, or have the whole field be fixed with MO equal and CDA not-sent. 1007 The same holds for the kid field. 1009 Figure 16 shows a possible set of Outer Rules to compress the Outer 1010 Header. 1012 Rule ID 0 1013 +-------------------+--+--+--------------+--------+---------++------+ 1014 | Field |FP|DI| Target | MO | CDA || Sent | 1015 | | | | Value | | ||[bits]| 1016 +-------------------+--+--+--------------+--------+---------++------+ 1017 |CoAP version | |bi| 01 |equal |not-sent || | 1018 |CoAP Type | |up| 0 |equal |not-sent || | 1019 |CoAP Type | |dw| 2 |equal |not-sent || | 1020 |CoAP TKL | |bi| 1 |equal |not-sent || | 1021 |CoAP Code | |up| 2 |equal |not-sent || | 1022 |CoAP Code | |dw| 68 |equal |not-sent || | 1023 |CoAP MID | |bi| 0000 |MSB(12) |LSB ||MMMM | 1024 |CoAP Token | |bi| 0x80 |MSB(5) |LSB ||TTT | 1025 |CoAP OSCORE_flags | |up| 0x09 |equal |not-sent || | 1026 |CoAP OSCORE_piv | |up| 0x00 |MSB(4) |LSB ||PPPP | 1027 |COAP OSCORE_kid | |up|0x636c69656e70|MSB(52) |LSB ||KKKK | 1028 |COAP OSCORE_kidctxt| |bi| b'' |equal |not-sent || | 1029 |CoAP OSCORE_flags | |dw| b'' |equal |not-sent || | 1030 |CoAP OSCORE_piv | |dw| b'' |equal |not-sent || | 1031 |CoAP OSCORE_kid | |dw| b'' |equal |not-sent || | 1032 |COAP Option-End | |dw| 0xFF |equal |not-sent || | 1033 +-------------------+--+--+--------------+--------+---------++------+ 1035 Figure 16: Outer SCHC Rules 1037 These Outer Rules are applied to the example GET Request and CONTENT 1038 Response. The resulting messages are shown in Figure 17 and 1039 Figure 18. 1041 Compressed message: 1042 ================== 1043 0x001489458a9fc3686852f6c4 (12 bytes) 1044 0x00 Rule ID 1045 1489 Compression Residue 1046 458a9fc3686852f6c4 Padded payload 1048 Compression residue: 1049 0b 0001 010 0100 0100 (15 bits -> 2 bytes with padding) 1050 mid tkn piv kid 1052 Payload 1053 0xa2c54fe1b434297b62 (9 bytes) 1055 Compressed message length: 12 bytes 1057 Figure 17: SCHC-OSCORE Compressed GET Request 1059 Compressed message: 1060 ================== 1061 0x0014218daf84d983d35de7e48c3c1852 (16 bytes) 1062 0x00 Rule ID 1063 14 Compression residue 1064 218daf84d983d35de7e48c3c1852 Padded payload 1065 Compression residue: 1066 0b0001 010 (7 bits -> 1 byte with padding) 1067 mid tkn 1069 Payload 1070 0x10c6d7c26cc1e9aef3f2461e0c29 (14 bytes) 1072 Compressed msg length: 16 bytes 1074 Figure 18: SCHC-OSCORE Compressed CONTENT Response 1076 For contrast, we compare these results with what would be obtained by 1077 SCHC compressing the original CoAP messages without protecting them 1078 with OSCORE. To do this, we compress the CoAP messages according to 1079 the SCHC rules in Figure 19. 1081 Rule ID 1 1082 +---------------+--+--+-----------+---------+-----------++--------+ 1083 | Field |FP|DI| Target | MO | CDA || Sent | 1084 | | | | Value | | || [bits] | 1085 +---------------+--+--+-----------+---------+-----------++--------+ 1086 |CoAP version | |bi| 01 |equal |not-sent || | 1087 |CoAP Type | |up| 0 |equal |not-sent || | 1088 |CoAP Type | |dw| 2 |equal |not-sent || | 1089 |CoAP TKL | |bi| 1 |equal |not-sent || | 1090 |CoAP Code | |up| 2 |equal |not-sent || | 1091 |CoAP Code | |dw| [69,132] |match-map|map-sent ||C | 1092 |CoAP MID | |bi| 0000 |MSB(12) |LSB ||MMMM | 1093 |CoAP Token | |bi| 0x80 |MSB(5) |LSB ||TTT | 1094 |CoAP Uri-Path | |up|temperature|equal |not-sent || | 1095 |COAP Option-End| |dw| 0xFF |equal |not-sent || | 1096 +---------------+--+--+-----------+---------+-----------++--------+ 1098 Figure 19: SCHC-CoAP Rules (No OSCORE) 1100 This yields the results in Figure 20 for the Request, and Figure 21 1101 for the Response. 1103 Compressed message: 1104 ================== 1105 0x0114 1106 0x01 = Rule ID 1108 Compression residue: 1109 0b00010100 (1 byte) 1111 Compressed msg length: 2 1113 Figure 20: CoAP GET Compressed without OSCORE 1115 Compressed message: 1116 ================== 1117 0x010a32332043 1118 0x01 = Rule ID 1120 Compression residue: 1121 0b00001010 (1 byte) 1123 Payload 1124 0x32332043 1126 Compressed msg length: 6 1128 Figure 21: CoAP CONTENT Compressed without OSCORE 1130 As can be seen, the difference between applying SCHC + OSCORE as 1131 compared to regular SCHC + COAP is about 10 bytes of cost. 1133 8. IANA Considerations 1135 This document has no request to IANA. 1137 9. Security considerations 1139 This document does not have any more Security consideration than the 1140 ones already raised on [I-D.ietf-lpwan-ipv6-static-context-hc] 1142 10. Acknowledgements 1144 The authors would like to thank Dominique Barthel, Carsten Bormann, 1145 Thomas Fossati, Klaus Hartke, Francesca Palombini, Alexander Pelov, 1146 Goran Selander. 1148 11. Normative References 1150 [I-D.ietf-lpwan-ipv6-static-context-hc] 1151 Minaburo, A., Toutain, L., Gomez, C., Barthel, D., and J. 1152 Zuniga, "Static Context Header Compression (SCHC) and 1153 fragmentation for LPWAN, application to UDP/IPv6", draft- 1154 ietf-lpwan-ipv6-static-context-hc-24 (work in progress), 1155 December 2019. 1157 [rfc2119] Bradner, S., "Key words for use in RFCs to Indicate 1158 Requirement Levels", BCP 14, RFC 2119, 1159 DOI 10.17487/RFC2119, March 1997, 1160 . 1162 [rfc7252] Shelby, Z., Hartke, K., and C. Bormann, "The Constrained 1163 Application Protocol (CoAP)", RFC 7252, 1164 DOI 10.17487/RFC7252, June 2014, 1165 . 1167 [rfc7641] Hartke, K., "Observing Resources in the Constrained 1168 Application Protocol (CoAP)", RFC 7641, 1169 DOI 10.17487/RFC7641, September 2015, 1170 . 1172 [rfc7959] Bormann, C. and Z. Shelby, Ed., "Block-Wise Transfers in 1173 the Constrained Application Protocol (CoAP)", RFC 7959, 1174 DOI 10.17487/RFC7959, August 2016, 1175 . 1177 [rfc7967] Bhattacharyya, A., Bandyopadhyay, S., Pal, A., and T. 1178 Bose, "Constrained Application Protocol (CoAP) Option for 1179 No Server Response", RFC 7967, DOI 10.17487/RFC7967, 1180 August 2016, . 1182 [rfc8174] Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC 1183 2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174, 1184 May 2017, . 1186 [rfc8613] Selander, G., Mattsson, J., Palombini, F., and L. Seitz, 1187 "Object Security for Constrained RESTful Environments 1188 (OSCORE)", RFC 8613, DOI 10.17487/RFC8613, July 2019, 1189 . 1191 Authors' Addresses 1193 Ana Minaburo 1194 Acklio 1195 1137A avenue des Champs Blancs 1196 35510 Cesson-Sevigne Cedex 1197 France 1199 Email: ana@ackl.io 1201 Laurent Toutain 1202 Institut MINES TELECOM; IMT Atlantique 1203 2 rue de la Chataigneraie 1204 CS 17607 1205 35576 Cesson-Sevigne Cedex 1206 France 1208 Email: Laurent.Toutain@imt-atlantique.fr 1209 Ricardo Andreasen 1210 Universidad de Buenos Aires 1211 Av. Paseo Colon 850 1212 C1063ACV Ciudad Autonoma de Buenos Aires 1213 Argentina 1215 Email: randreasen@fi.uba.ar