idnits 2.17.1 draft-perlman-trill-rbridge-multilevel-09.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 : ---------------------------------------------------------------------------- == There are 1 instance of lines with non-RFC6890-compliant IPv4 addresses in the document. If these are example addresses, they should be changed. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (March 5, 2015) is 3334 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- ** Obsolete normative reference: RFC 6439 (Obsoleted by RFC 8139) Summary: 1 error (**), 0 flaws (~~), 2 warnings (==), 1 comment (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 1 TRILL Working Group Radia Perlman 2 INTERNET-DRAFT EMC 3 Intended status: Informational Donald Eastlake 4 Mingui Zhang 5 Huawei 6 Anoop Ghanwani 7 Dell 8 Hongjun Zhai 9 ZTE 10 Expires: September 4, 2015 March 5, 2015 12 Flexible Multilevel TRILL 13 (Transparent Interconnection of Lots of Links) 14 16 Abstract 18 Extending TRILL to multiple levels has challenges that are not 19 addressed by the already-existing capability of IS-IS to have 20 multiple levels. One issue is with the handling of multi-destination 21 packet distribution trees. Another issue is with TRILL switch 22 nicknames. There have been two proposed approaches. One approach, 23 which we refer to as the "unique nickname" approach, gives unique 24 nicknames to all the TRILL switches in the multilevel campus, either 25 by having the level-1/level-2 border TRILL switches advertise which 26 nicknames are not available for assignment in the area, or by 27 partitioning the 16-bit nickname into an "area" field and a "nickname 28 inside the area" field. The other approach, which we refer to as the 29 "aggregated nickname" approach, involves hiding the nicknames within 30 areas, allowing nicknames to be reused in different areas, by having 31 the border TRILL switches rewrite the nickname fields when entering 32 or leaving an area. Each of those approaches has advantages and 33 disadvantages. This informational document suggests allowing a choice 34 of approach in each area. This allows the simplicity of the unique 35 nickname approach in installations in which there is no danger of 36 running out of nicknames and allows the complexity of hiding the 37 nicknames in an area to be phased into larger installations on a per- 38 area basis. 40 Status of This Memo 42 This Internet-Draft is submitted to IETF in full conformance with the 43 provisions of BCP 78 and BCP 79. Distribution of this document is 44 unlimited. Comments should be sent to the TRILL working group 45 mailing list . 47 Internet-Drafts are working documents of the Internet Engineering 48 Task Force (IETF), its areas, and its working groups. Note that 49 other groups may also distribute working documents as Internet- 50 Drafts. 52 Internet-Drafts are draft documents valid for a maximum of six months 53 and may be updated, replaced, or obsoleted by other documents at any 54 time. It is inappropriate to use Internet-Drafts as reference 55 material or to cite them other than as "work in progress." 57 The list of current Internet-Drafts can be accessed at 58 http://www.ietf.org/1id-abstracts.html. The list of Internet-Draft 59 Shadow Directories can be accessed at 60 http://www.ietf.org/shadow.html. 62 Table of Contents 64 1. Introduction............................................4 65 1.1 TRILL Scalability Issues...............................4 66 1.2 Improvements Due to Multilevel.........................5 67 1.3 Unique and Aggregated Nickanmes........................6 68 1.3 More on Areas..........................................6 69 1.4 Terminology and Acronyms...............................7 71 2. Multilevel TRILL Issues.................................8 72 2.1 Non-zero Area Addresses................................9 73 2.2 Aggregated versus Unique Nicknames.....................9 74 2.2.1 More Details on Unique Nicknames....................10 75 2.2.2 More Details on Aggregated Nicknames................11 76 2.2.2.1 Border Learning Aggregated Nicknames..............12 77 2.2.2.2 Swap Nickname Field Aggregated Nicknames..........14 78 2.2.2.3 Comparison........................................14 79 2.3 Building Multi-Area Trees.............................15 80 2.4 The RPF Check for Trees...............................15 81 2.5 Area Nickname Acquisition.............................15 82 2.6 Link State Representation of Areas....................16 84 3. Area Partition.........................................18 86 4. Multi-Destination Scope................................19 87 4.1 Unicast to Multi-destination Conversions..............19 88 4.1.1 New Tree Encoding...................................20 89 4.2 Selective Broadcast Domain Reduction..................20 91 5. Co-Existence with Old TRILL switches...................22 92 6. Multi-Access Links with End Stations...................23 93 7. Summary................................................24 95 8. Security Considerations................................25 96 9. IANA Considerations....................................25 98 Normative References......................................26 99 Informative References....................................26 100 Acknowledgements..........................................28 101 Authors' Addresses........................................29 103 1. Introduction 105 The IETF TRILL (Transparent Interconnection of Lot of Links or 106 Tunneled Routing in the Link Layer) protocol [RFC6325] [RFC7177] 107 provides optimal pair-wise data routing without configuration, safe 108 forwarding even during periods of temporary loops, and support for 109 multipathing of both unicast and multicast traffic in networks with 110 arbitrary topology and link technology, including multi-access links. 111 TRILL accomplishes this by using IS-IS (Intermediate System to 112 Intermediate System [IS-IS] [RFC7176]) link state routing in 113 conjunction with a header that includes a hop count. The design 114 supports data labels (VLANs and Fine Grained Labels [RFC7172]) and 115 optimization of the distribution of multi-destination data based on 116 VLANs and multicast groups. Devices that implement TRILL are called 117 TRILL Switches or RBridges. 119 Familiarity with [RFC6325] and [rfc7180bis] is assumed in this 120 document. 122 1.1 TRILL Scalability Issues 124 There are multiple issues that might limit the scalability of a 125 TRILL-based network: 127 1. the routing computation load, 128 2. the volatility of the link state database (LSDB) creating too much 129 control traffic, 130 3. the volatility of the LSDB causing the TRILL network to be in an 131 unconverged state too much of the time, 132 4. the size of the LSDB, 133 5. the limit of the number of TRILL switches, due to the 16-bit 134 nickname space, 135 6. the traffic due to upper layer protocols use of broadcast and 136 multicast, and 137 7. the size of the end node learning table (the table that remembers 138 (egress TRILL switch, label/MAC) pairs). 140 Extending TRILL IS-IS to be multilevel (hierarchical) helps with all 141 but the last of these issues. 143 IS-IS was designed to be multilevel [IS-IS]. A network can be 144 partitioned into "areas". Routing within an area is known as "Level 145 1 routing". Routing between areas is known as "Level 2 routing". 146 The Level 2 IS-IS network consists of Level 2 routers and links 147 between the Level 2 routers. Level 2 routers may participate in one 148 or more Level 1 areas, in addition to their role as Level 2 routers. 150 Each area is connected to Level 2 through one or more "border 151 routers", which participate both as a router inside the area, and as 152 a router inside the Level 2 "area". Care must be taken that it is 153 clear, when transitioning multi-destination packets between Level 2 154 and a Level 1 area in either direction, that exactly one border TRILL 155 switch will transition a particular data packet between the levels or 156 else duplication or loss of traffic can occur. 158 1.2 Improvements Due to Multilevel 160 Partitioning the network into areas solves the first four scalability 161 issues described above, namely, 163 1. the routing computation load, 165 2. the volatility of the LSDB creating too much control traffic, 167 3. the volatility of the LSDB causing the TRILL network to be in an 168 unconverged state too much of the time, 170 4. the size of the LSDB. 172 Problem #6 in Section 1.1, namely, the traffic due to upper layer 173 protocols use of broadcast and multicast, can be addressed by 174 introducing a locally-scoped multi-destination delivery, limited to 175 an area or a single link. See further discussion in Section 4.2. 177 Problem #5 in Section 1.1, namely, the limit of the number of TRILL 178 switches, due to the 16-bit nickname space, will only be addressed 179 with the aggregated nickname approach. Since the aggregated nickname 180 approach requires some complexity in the border TRILL switches (for 181 rewriting the nicknames in the TRILL header), the design in this 182 document allows a campus with a mixture of unique-nickname areas, and 183 aggregated-nickname areas. Nicknames must be unique across all Level 184 2 and unique-nickname area TRILL switches, whereas nicknames inside 185 an aggregated-nickname area are visible only inside the area. 186 Nicknames inside an aggregated-nickname area must not conflict with 187 nicknames visible in Level 2 (which includes all nicknames inside 188 unique nickname areas), but the nicknames inside an aggregated- 189 nickname area may be the same as nicknames used within other 190 aggregated-nickname areas. 192 TRILL switches within an area need not be aware of whether they are 193 in an aggregated nickname area or a unique nickname area. The border 194 TRILL switches in area A1 will claim, in their LSP inside area A1, 195 which nicknames (or nickname ranges) are not available for choosing 196 as nicknames by area A1 TRILL switches. 198 1.3 Unique and Aggregated Nickanmes 200 We describe two alternatives for hierarchical or multilevel TRILL. 201 One we call the "unique nickname" alternative. The other we call the 202 "aggregated nickname" alternative. In the aggregated nickname 203 alternative, border TRILL switches replace either the ingress or 204 egress nickname field in the TRILL header of unicast packets with an 205 aggregated nickname representing an entire area. 207 The unique nickname alternative has the advantage that border TRILL 208 switches are simpler and do not need to do TRILL Header nickname 209 modification. It also simplifies testing and maintenance operations 210 that originate in one area and terminate in a different area. 212 The aggregated nickname alternative has the following advantages: 214 o it solves problem #5 above, the 16-bit nickname limit, in a 215 simple way, 216 o it lessens the amount of inter-area routing information that 217 must be passed in IS-IS, and 218 o it logically reduces the RPF (Reverse Path Forwarding) Check 219 information (since only the area nickname needs to appear, 220 rather than all the ingress TRILL switches in that area). 222 In both cases, it is possible and advantageous to compute multi- 223 destination data packet distribution trees such that the portion 224 computed within a given area is rooted within that area. 226 1.3 More on Areas 228 Each area is configured with an "area address", which is advertised 229 in IS-IS messages, so as to avoid accidentally interconnecting areas. 230 Although the area address had other purposes in CLNP (IS-IS was 231 originally designed for CLNP/DECnet), for TRILL the only purpose of 232 the area address would be to avoid accidentally interconnecting 233 areas. 235 Currently, the TRILL specification says that the area address must be 236 zero. If we change the specification so that the area address value 237 of zero is just a default, then most of IS-IS multilevel machinery 238 works as originally designed. However, there are TRILL-specific 239 issues, which we address below in this document. 241 1.4 Terminology and Acronyms 243 This document generally uses the acronyms defined in [RFC6325] plus 244 the additional acronym DBRB. However, for ease of reference, most 245 acronyms used are listed here: 247 CLNP - ConnectionLess Network Protocol 249 DECnet - a proprietary routing protocol that was used by Digital 250 Equipment Corporation. "DECnet Phase 5" was the origin of IS-IS. 252 Data Label - VLAN or Fine Grained Label [RFC7172] 254 DBRB - Designated Border RBridge 256 IS-IS - Intermediate System to Intermediate System [IS-IS] 258 LSDB - Link State Data Base 260 LSP - Link Stat PDU 262 PDU - Protocol Data Unit 264 RBridge - Routing Bridge, an alterntive name for a TRILL switch 266 RPF - Reverse Path Forwarding 268 TRILL - Transparent Interconnection of Lots of Links or Tunneled 269 Routing in the Link Layer [RFC6325] 271 TRILL switch - an alternative name for an RBridge 273 VLAN - Virtual Local Area Network 275 2. Multilevel TRILL Issues 277 The TRILL-specific issues introduced by multilevel include the 278 following: 280 a. Configuration of non-zero area addresses, encoding them in IS-IS 281 PDUs, and possibly interworking with old TRILL switches that do 282 not understand nonzero area addresses. 284 See Section 2.1. 286 b. Nickname management. 288 See Sections 2.5 and 2.2. 290 c. Advertisement of pruning information (Data Label reachability, IP 291 multicast addresses) across areas. 293 Distribution tree pruning information is only an optimization, 294 as long as multi-destination packets are not prematurely 295 pruned. For instance, border TRILL switches could advertise 296 they can reach all possible Data Labels, and have an IP 297 multicast router attached. This would cause all multi- 298 destination traffic to be transmitted to border TRILL switches, 299 and possibly pruned there, when the traffic could have been 300 pruned earlier based on Data Label or multicast group if border 301 TRILL switches advertised more detailed Data Label and/or 302 multicast listener and multicast router attachment information. 304 d. Computation of distribution trees across areas for multi- 305 destination data. 307 See Section 2.3. 309 e. Computation of RPF information for those distribution trees. 311 See Section 2.4. 313 f. Computation of pruning information across areas. 315 See Sections 2.3 and 2.6. 317 g. Compatibility, as much as practical, with existing, unmodified 318 TRILL switches. 320 The most important form of compatibility is with existing TRILL 321 fast path hardware. Changes that require upgrade to the slow 322 path firmware/software are more tolerable. Compatibility for 323 the relatively small number of border TRILL switches is less 324 important than compatibility for non-border TRILL switches. 326 See Section 5. 328 2.1 Non-zero Area Addresses 330 The current TRILL base protocol specification [RFC6325] [RFC7177] 331 [rfc7180bis] says that the area address in IS-IS must be zero. The 332 purpose of the area address is to ensure that different areas are not 333 accidentally merged. Furthermore, zero is an invalid area address 334 for layer 3 IS-IS, so it was chosen as an additional safety mechanism 335 to ensure that layer 3 IS-IS would not be confused with TRILL IS-IS. 336 However, TRILL uses other techniques to avoid such confusion, such as 337 different multicast addresses and Ethertypes on Ethernet [RFC6325], 338 different PPP codepoints on PPP [RFC6361], and the the like, so use 339 in TRILL of an area address that might be used in layer 3 IS-IS is 340 not a problem. 342 Since current TRILL switches will reject any IS-IS messages with 343 nonzero area addresses, the choices are as follows: 345 a.1 upgrade all TRILL switches that are to interoperate in a 346 potentially multilevel environment to understand non-zero area 347 addresses, 348 a.2 neighbors of old TRILL switches must remove the area address from 349 IS-IS messages when talking to an old TRILL switch (which might 350 break IS-IS security and/or cause inadvertent merging of areas), 351 a.3 ignore the problem of accidentally merging areas entirely, or 352 a.4 keep the fixed "area address" field as 0 in TRILL, and add a new, 353 optional TLV for "area name" that, if present, could be compared, 354 by new TRILL switches, to prevent accidental area merging. 356 In principal, different solutions could be used in different areas 357 but it would be much simpler to adopt one of these choices uniformly. 359 2.2 Aggregated versus Unique Nicknames 361 In the unique nickname alternative, all nicknames across the campus 362 must be unique. In the aggregated nickname alternative, TRILL switch 363 nicknames within an aggregated area are only of local significance, 364 and the only nickname externally (outside that area) visible is the 365 "area nickname" (or nicknames), which aggregates all the internal 366 nicknames. 368 The unique nickname approach simplifies border TRILL switches. 370 The aggregated nickname approach eliminates the potential problem of 371 nickname exhaustion, minimizes the amount of nickname information 372 that would need to be forwarded between areas, minimizes the size of 373 the forwarding table, and simplifies RPF calculation and RPF 374 information. 376 2.2.1 More Details on Unique Nicknames 378 With unique cross-area nicknames, it would be intractable to have a 379 flat nickname space with TRILL switches in different areas contending 380 for the same nicknames. Instead, each area would need to be 381 configured with a block of nicknames. Either some TRILL switches 382 would need to announce that all the nicknames other than that block 383 are taken (to prevent the TRILL switches inside the area from 384 choosing nicknames outside the area's nickname block), or a new TLV 385 would be needed to announce the allowable nicknames, and all TRILL 386 switches in the area would need to understand that new TLV. An 387 example of the second approach is given in [NickFlags]. 389 Currently the encoding of nickname information in TLVs is by listing 390 of individual nicknames; this would make it painful for a border 391 TRILL switch to announce into an area that it is holding all other 392 nicknames to limit the nicknames available within that area. The 393 information could be encoded as ranges of nicknames to make this 394 somewhat manageable [NickFlags]; however, a new TLV for announcing 395 nickname ranges would not be intelligible to old TRILL switches. 397 There is also an issue with the unique nicknames approach in building 398 distribution trees, as follows: 400 With unique nicknames in the TRILL campus and TRILL header 401 nicknames not rewritten by the border TRILL switches, there would 402 have to be globally known nicknames for the trees. Suppose there 403 are k trees. For all of the trees with nicknames located outside 404 an area, the local trees would be rooted at a border TRILL switch 405 or switches. Therefore, there would be either no splitting of 406 multi-destination traffic with the area or restricted splitting of 407 multi-destination traffic between trees rooted at a highly 408 restricted set of TRILL switches. 410 As an alternative, just the "egress nickname" field of multi- 411 destination TRILL Data packets could be mapped at the border, 412 leaving known unicast packets un-mapped. However, this surrenders 413 much of the unique nickname advantage of simpler border TRILL 414 switches. 416 Scaling to a very large campus with unique nicknames might exhaust 417 the 16-bit TRILL nicknames space. One method might be to expand 418 nicknames to 24bits; however, that technique would require TRILL 419 message format changes and that all TRILL switches in the campus 420 understand larger nicknames. 422 For an example of a more specific multilevel proposal using unique 423 nicknames, see [DraftUnique]. 425 2.2.2 More Details on Aggregated Nicknames 427 The aggregated nickname approach enables passing far less nickname 428 information. It works as follows, assuming both the source and 429 destination areas are using aggregated nicknames: 431 Each area would be assigned a 16-bit nickname. This would not be 432 the nickname of any actual TRILL switch. Instead, it would be the 433 nickname of the area itself. Border TRILL switches would know the 434 area nickname for their own area(s). 436 The TRILL Header nickname fields in TRILL Data packets being 437 transported through a multilevel TRILL campus with aggregated 438 nicknames are as follows: 440 - When both the ingress and egress TRILL switches are in the same 441 area, there need be no change from the existing base TRILL 442 protocol standard in the TRILL Header nickname fields. 444 - When being transported in Level 2, the ingress nickname is the 445 nickname of the ingress TRILL switch's area while the egress 446 nickname is either the nickname of the egress TRILL switch's 447 area or a tree nickname. 449 - When being transported from Level 1 to Level 2, the ingress 450 nickname is the nickname of the ingress TRILL switch itself 451 while the egress nickname is either the nickname of the area of 452 the egress TRILL switch or a tree nickname. 454 - When being transported from Level 2 to Level 1, the ingress 455 nickname is the nickname of the ingress TRILL switch's area 456 while the egress nickname is either the nickname of the egress 457 TRILL switch itself or a tree nickname. 459 There are two variations of the aggregated nickname approach. The 460 first is the Border Learning approach, which is described in Section 461 2.2.2.1. The second is the Swap Nickname Field approach, which is 462 described in Section 2.2.2.2. Section 2.2.2.3 compares the advantages 463 and disadvantages of these two variations of the aggregated nickname 464 approach. 466 2.2.2.1 Border Learning Aggregated Nicknames 468 This section provides an illustrative example and description of the 469 border learning variation of aggregated nicknames. 471 In the following picture, RB2 and RB3 are area border TRILL switches 472 (RBridges). A source S is attached to RB1. The two areas have 473 nicknames 15961 and 15918, respectively. RB1 has a nickname, say 27, 474 and RB4 has a nickname, say 44 (and in fact, they could even have the 475 same nickname, since the TRILL switch nickname will not be visible 476 outside these aggreated areas). 478 Area 15961 level 2 Area 15918 479 +-------------------+ +-----------------+ +--------------+ 480 | | | | | | 481 | S--RB1---Rx--Rz----RB2---Rb---Rc--Rd---Re--RB3---Rk--RB4---D | 482 | 27 | | | | 44 | 483 | | | | | | 484 +-------------------+ +-----------------+ +--------------+ 486 Let's say that S transmits a frame to destination D, which is 487 connected to RB4, and let's say that D's location has already been 488 learned by the relevant TRILL switches. These relevant switches have 489 learned the following: 491 1) RB1 has learned that D is connected to nickname 15918 492 2) RB3 has learned that D is attached to nickname 44. 494 The following sequence of events will occur: 496 - S transmits an Ethernet frame with source MAC = S and destination 497 MAC = D. 499 - RB1 encapsulates with a TRILL header with ingress RBridge = 27, 500 and egress = 15918 producing a TRILL Data packet. 502 - RB2 has announced in the Level 1 IS-IS instance in area 15961, 503 that it is attached to all the area nicknames, including 15918. 504 Therefore, IS-IS routes the packet to RB2. Alternatively, if a 505 distinguished range of nicknames is used for Level 2, Level 1 506 TRILL switches seeing such an egress nickname will know to route 507 to the nearest border router, which can be indicated by the IS-IS 508 attached bit. 510 - RB2, when transitioning the packet from Level 1 to Level 2, 511 replaces the ingress TRILL switch nickname with the area nickname, 512 so replaces 27 with 15961. Within Level 2, the ingress RBridge 513 field in the TRILL header will therefore be 15961, and the egress 514 RBridge field will be 15918. Also RB2 learns that S is attached to 515 nickname 27 in area 15961 to accommodate return traffic. 517 - The packet is forwarded through Level 2, to RB3, which has 518 advertised, in Level 2, reachability to the nickname 15918. 520 - RB3, when forwarding into area 15918, replaces the egress nickname 521 in the TRILL header with RB4's nickname (44). So, within the 522 destination area, the ingress nickname will be 15961 and the 523 egress nickname will be 44. 525 - RB4, when decapsulating, learns that S is attached to nickname 526 15961, which is the area nickname of the ingress. 528 Now suppose that D's location has not been learned by RB1 and/or RB3. 529 What will happen, as it would in TRILL today, is that RB1 will 530 forward the packet as multi-destination, choosing a tree. As the 531 multi-destination packet transitions into Level 2, RB2 replaces the 532 ingress nickname with the area nickname. If RB1 does not know the 533 location of D, the packet must be flooded, subject to possible 534 pruning, in Level 2 and, subject to possible pruning, from Level 2 535 into every Level 1 area that it reaches on the Level 2 distribution 536 tree. 538 Now suppose that RB1 has learned the location of D (attached to 539 nickname 15918), but RB3 does not know where D is. In that case, RB3 540 must turn the packet into a multi-destination packet within area 541 15918. In this case, care must be taken so that, in case RB3 is not 542 the Designated transitioner between Level 2 and its area for that 543 multi-destination packet, but was on the unicast path, that another 544 border TRILL switch in that area not forward the now multi- 545 destination packet back into Level 2. Therefore, it would be 546 desirable to have a marking, somehow, that indicates the scope of 547 this packet's distribution to be "only this area" (see also Section 548 4). 550 In cases where there are multiple transitioners for unicast packets, 551 the border learning mode of operation requires that the address 552 learning between them be shared by some protocol such as running 553 ESADI [RFC7357] for all Data Labels of interest to avoid excessive 554 unknown unicast flooding. 556 The potential issue described at the end of Section 2.2.1 with trees 557 in the unique nickname alternative is eliminated with aggregated 558 nicknames. With aggregated nicknames, each border TRILL switch that 559 will transition multi-destination packets can have a mapping between 560 Level 2 tree nicknames and Level 1 tree nicknames. There need not 561 even be agreement about the total number of trees; just that the 562 border TRILL switch have some mapping, and replace the egress TRILL 563 switch nickname (the tree name) when transitioning levels. 565 2.2.2.2 Swap Nickname Field Aggregated Nicknames 567 As a variant, two additional fields could exist in TRILL Data packets 568 we call the "ingress swap nickname field" and the "egress swap 569 nickname field". The changes in the example above would be as 570 follows: 572 - RB1 will have learned the area nickname of D and the TRILL switch 573 nickname of RB4 to which D is attached. In encapsulating a frame 574 to D, it puts the area nickname of D (15918) in the egress 575 nickname field of the TRILL Header and puts the nickname of RB3 576 (44) in a egress swap nickname field. 578 - RB2 moves the ingress nickname to the ingress swap nickname field 579 and inserts 15961, the area nickname for S, into the ingress 580 nickname field. 582 - RB3 swaps the egress nickname and the egress swap nickname fields, 583 which sets the egress nickname to 44. 585 - RB4 learns the correspondence between the source MAC/VLAN of S and 586 the { ingress nickname, ingress swap nickname field } pair as it 587 decapsulates and egresses the frame. 589 See [DraftAggregated] for a multilevel proposal using aggregated swap 590 nicknames. 592 2.2.2.3 Comparison 594 The Border Learning variant described in Section 2.2.2.1 above 595 minimizes the change in non-border TRILL switches but imposes the 596 burden on border TRILL switches of learning and doing lookups in all 597 the end station MAC addresses within their area(s) that are used for 598 communication outside the area. This burden could be reduced by 599 decreasing the area size and increasing the number of areas. 601 The Swap Nickname Field variant described in Section 2.2.2.2 602 eliminates the extra address learning burden on border TRILL switches 603 but requires more extensive changes to non-border TRILL switches. In 604 particular they must learn to associate both a TRILL switch nickname 605 and an area nickname with end station MAC/label pairs (except for 606 addresses that are local to their area). 608 The Swap Nickname Field alternative is more scalable but less 609 backward compatible for non-border TRILL switches. It would be 610 possible for border and other level 2 TRILL switches to support both 611 Border Learning, for support of legacy Level 1 TRILL switches, and 612 Swap Nickname, to support Level 1 TRILL switches that understood the 613 Swap Nickname method. 615 2.3 Building Multi-Area Trees 617 It is easy to build a multi-area tree by building a tree in each area 618 separately, (including the Level 2 "area"), and then having only a 619 single border TRILL switch, say RBx, in each area, attach to the 620 Level 2 area. RBx would forward all multi-destination packets 621 between that area and Level 2. 623 People might find this unacceptable, however, because of the desire 624 to path split (not always sending all multi-destination traffic 625 through the same border TRILL switch). 627 This is the same issue as with multiple ingress TRILL switches 628 injecting traffic from a pseudonode, and can be solved with the 629 mechanism that was adopted for that purpose: the affinity TLV 630 [DraftCMT]. For each tree in the area, at most one border RB 631 announces itself in an affinity TLV with that tree name. 633 2.4 The RPF Check for Trees 635 For multi-destination data originating locally in RBx's area, 636 computation of the RPF check is done as today. For multi-destination 637 packets originating outside RB1's area, computation of the RPF check 638 must be done based on which one of the border TRILL switches (say 639 RB1, RB2, or RB3) injected the packet into the area. 641 A TRILL switch, say RB4, located inside an area, must be able to know 642 which of RB1, RB2, or RB3 transitioned the packet into the area from 643 Level 2. (or into Level 2 from an area). 645 This could be done based on having the DBRB announce the transitioner 646 assignments to all the TRILL switches in the area, or the Affinity 647 TLV mechanism given in [DraftCMT], or the New Tree Encoding mechanism 648 discussed in Section 4.1.1. 650 2.5 Area Nickname Acquisition 652 In the aggregated nickname alternative, each area must acquire a 653 unique area nickname. It is probably simpler to allocate a block of 654 nicknames (say, the top 4000) to be area addresses, and not used by 655 any TRILL switches. 657 The area nicknames need to be advertised and acquired through Level 658 2. 660 Within an area, all the border TRILL switches must discover each 661 other through the Level 1 link state database, by using the IS-IS 662 attach bit or by explicitly advertising in their LSP "I am a border 663 RBridge". 665 Of the border TRILL switches, one will have highest priority (say 666 RB7). RB7 can dynamically participate, in Level 2, to acquire a 667 pseudo-nickname for the area analagous to the pseudo-nickname for an 668 active-active edge group [PseudoNickname]. Alternatively, RB7 could 669 give the area a pseudonode IS-IS ID, such as RB7.5, within Level 2. 670 So an area would appear, in Level 2, as a pseudonode and the 671 pseudonode can participate, in Level 2, to acquire a nickname for the 672 area. 674 Within Level 2, all the border TRILL switches for an area can 675 advertise reachability to the area, which would mean connectivity to 676 the area nickname. 678 2.6 Link State Representation of Areas 680 Within an area, say area A1, there is an election for the DBRB, 681 (Designated Border RBridge), say RB1. This can be done through LSPs 682 within area A1. The border TRILL switches announce themselves, 683 together with their DBRB priority. (Note that the election of the 684 DBRB cannot be done based on Hello messages, because the border TRILL 685 switches are not necessarily physical neighbors of each other. They 686 can, however, reach each other through connectivity within the area, 687 which is why it will work to find each other through Level 1 LSPs.) 689 RB1 acquires the area nickname (in the aggregated nickname approach) 690 and may give the area a pseudonode IS-IS ID (just like the DRB would 691 give a pseudonode IS-IS ID to a link) depending on how the area 692 nickname is handled. RB1 advertises, in area A1, the area nickname 693 that RB1 has acquired (and what the pseudonode IS-IS ID for the area 694 is if needed). 696 Level 1 LSPs (possibly pseudonode) initiated by RB1 for the area 697 include any information external to area A1 that should be input into 698 area A1 (such as area nicknames of external areas, or perhaps (in the 699 unique nickname variant) all the nicknames of external TRILL switches 700 in the TRILL campus and pruning information such as multicast 701 listeners and labels). All the other border TRILL switches for the 702 area announce (in their LSP) attachment to that area. 704 Within Level 2, RB1 generates a Level 2 LSP on behalf of the area. 706 The same pseudonode ID could be used within Level 1 and Level 2, for 707 the area. (There does not seem any reason why it would be useful for 708 it to be different, but there's also no reason why it would need to 709 be the same). Likewise, all the area A1 border TRILL switches would 710 announce, in their Level 2 LSPs, connection to the area. 712 3. Area Partition 714 It is possible for an area to become partitioned, so that there is 715 still a path from one section of the area to the other, but that path 716 is via the Level 2 area. 718 With multilevel TRILL, an area will naturally break into two areas in 719 this case. 721 Area addresses might be configured to ensure two areas are not 722 inadvertently connected. Area addresses appears in Hellos and LSPs 723 within the area. If two chunks, connected only via Level 2, were 724 configured with the same area address, this would not cause any 725 problems. (They would just operate as separate Level 1 areas.) 727 A more serious problem occurs if the Level 2 area is partitioned in 728 such a way that it could be healed by using a path through a Level 1 729 area. TRILL will not attempt to solve this problem. Within the Level 730 1 area, a single border RBridge will be the DBRB, and will be in 731 charge of deciding which (single) RBridge will transition any 732 particular multi-destination packets between that area and Level 2. 733 If the Level 2 area is partitioned, this will result in multi- 734 destination data only reaching the portion of the TRILL campus 735 reachable through the partition attached to the TRILL switch that 736 transitions that packet. It will not cause a loop. 738 4. Multi-Destination Scope 740 There are at least two reasons it would be desirable to be able to 741 mark a multi-destination packet with a scope that indicates the 742 packet should not exit the area, as follows: 744 1. To address an issue in the border learning variant of the 745 aggregated nickname alternative, when a unicast packet turns into 746 a multi-destination packet when transitioning from Level 2 to 747 Level 1, as discussed in Section 4.1. 749 2. To constrain the broadcast domain for certain discovery, 750 directory, or service protocols as discussed in Section 4.2. 752 Multi-destination packet distribution scope restriction could be done 753 in a number of ways. For example, there could be a flag in the packet 754 that means "for this area only". However, the technique that might 755 require the least change to TRILL switch fast path logic would be to 756 indicate this in the egress nickname that designates the distribution 757 tree being used. There could be two general tree nicknames for each 758 tree, one being for distribution restricted to the area and the other 759 being for multi-area trees. Or there would be a set of N (perhaps 16) 760 special currently reserved nicknames used to specify the N highest 761 priority trees but with the variation that if the special nickname is 762 used for the tree, the packet is not transitioned between areas. Or 763 one or more special trees could be built that were restricted to the 764 local area. 766 4.1 Unicast to Multi-destination Conversions 768 In the border learning variant of the aggregated nickname 769 alternative, a unicast packet might be known at the Level 1 to Level 770 2 transition, be forwarded as a unicast packet to the least cost 771 border TRILL switch advertising connectivity to the destination area, 772 but turn out to have an unknown destination { MAC, Data Label } pair 773 when it arrives at that border TRILL switch. 775 In this case, the packet must be converted into a multi-destination 776 packet and flooded in the destination area. However, if the border 777 TRILL switch doing the conversion is not the border TRILL switch 778 designated to transition the resulting multi-destination packet, 779 there is the danger that the designated transitioner may pick up the 780 packet and flood it back into Level 2 from which it may be flooded 781 into multiple areas. This danger can be avoided by restricting any 782 multi-destination packet that results from such a conversion to the 783 destination area through a flag in the packet or though distributing 784 it on a tree that is restricted to the area, or other techniques (see 785 Section 4). 787 Alternatively, a multi-destination packet intended only for the area 788 could be tunneled (within the area) to the RBridge RBx, that is the 789 appointed transitioner for that form of packet (say, based on VLAN or 790 FGL), with instructions that RBx only transmit the packet within the 791 area, and RBx could initiate the multi-destination packet within the 792 area. Since RBx introduced the packet, and is the only one allowed 793 to transition that packet to Level 2, this would accomplish scoping 794 of the packet to within the area. Since this case only occurs in the 795 unusual case when unicast packets need to be turned into multi- 796 destination as described above, the suboptimality of tunneling 797 between the border TRILL switch that receives the unicast packet and 798 the appointed level transitioner for that packet, would not be an 799 issue. 801 4.1.1 New Tree Encoding 803 The current encoding, in a TRILL header, of a tree, is of the 804 nickname of the tree root. This requires all 16 bits of the egress 805 nickname field. TRILL could instead, for example, use the bottom 6 806 bits to encode the tree number (allowing 64 trees), leavinig 10 bits 807 to encode information such as: 809 o scope: a flag indicating whether it should be single area only, or 810 entire campus 811 o border injector: an indicator of which of the k border TRILL 812 switches injected this packet 814 If TRILL were to adopt this new encoding, it would also avoid the 815 limitations of the Affinity sub-TLV [DraftCMT] in the single area 816 case [PseudoNickname]; any of the TRILL switches in an edge group 817 could inject a multi-destination packet. This would require all TRILL 818 switches to be changed to understand the new encoding for a tree, and 819 it would require a TLV in the LSP to indicate which number each of 820 the TRILL switches in an edge group would be. 822 4.2 Selective Broadcast Domain Reduction 824 There are a number of service, discovery, and directory protocols 825 that, for convenience, are accessed via multicast or broadcast 826 frames. Examples are DHCP, the NetBIOS Service Location Protocol, and 827 multicast DNS. 829 Some such protocols provide means to restrict distribution to an IP 830 subnet or equivalent to reduce size of the broadcast domain they are 831 using and then provide a proxy that can be placed in that subnet to 832 use unicast to access a service elsewhere. In cases where a proxy 833 mechanism is not currently defined, it may be possible to create one 834 that references a central server or cache. With multilevel TRILL, it 835 is possible to construct very large IP subnets that could become 836 saturated with multi-destination traffic of this type unless packets 837 can be further restricted in their distribution. Such restricted 838 distribution can be accomplished for some protocols, say protocol P, 839 in a variety of waying including the following: 841 - Either (1) at all ingress TRILL switches in an area place all 842 protocol P multi-destination packets on a distribution tree in 843 such a way that the packets are restricted to the area or (2) at 844 all border TRILL switches between that area and Level 2, detect 845 protocol P multi-destination packets and do not transition them. 847 - Then place one, or a few for redundancy, protocol P proxyies 848 inside each area where protocol P may be in use. These proxies 849 unicast protocol P requests or other messages to the actual campus 850 server(s) for P. They also receive unicast responses or other 851 messages from those servers and deliver them within the area via 852 unicast, multicast, or broadcast as appropriate. (Such proxies 853 would not be needed if it was acceptable for all protocol P 854 traffic to be restricted to an area.) 856 While it might seem logical to connect the campus servers to TRILL 857 switches in Level 2, they could be placed within one or more areas so 858 that, in some cases, those areas might not require a local proxy 859 server. 861 5. Co-Existence with Old TRILL switches 863 TRILL switches that are not multilevel aware may have a problem with 864 calculating RPF Check and filtering information, since they would not 865 be aware of assignment of border TRILL switch transitioning. 867 A possible solution, as long as any old TRILL switches exist within 868 an area, is to have the border TRILL switches elect a single DBRB 869 (Designated Border RBridge), and have all inter-area traffic go 870 through the DBRB (unicast as well as multi-destination). If that 871 DBRB goes down, a new one will be elected, but at any one time, all 872 inter-area traffic (unicast as well as multi-destination) would go 873 through that one DRBR. However this eliminates load splitting at 874 level transition. 876 6. Multi-Access Links with End Stations 878 Care must be taken, in the case where there are multiple TRILL 879 switches on a link with end stations, that only one TRILL switch 880 ingress/egress any given data packet from/to the end nodes. With 881 existing, single level TRILL, this is done by electing a single 882 Designated RBridge per link, which appoints a single Appointed 883 Forwarder per VLAN [RFC7177] [RFC6439]. But suppose there are two 884 (or more) TRILL switches on a link in different areas, say RB1 in 885 area 1000 and RB2 in area 2000, and that the link contains end nodes. 886 If RB1 and RB2 ignore each other's Hellos then they will both 887 ingress/egress end node traffic from the link. 889 A simple rule is to use the TRILL switch or switches having the 890 lowest numbered area, comparing area numbers as unsigned integers, to 891 handle native traffic. This would automatically give multilevel- 892 ignorant legacy TRILL switches, that would be using area number zero, 893 highest priority for handling end stations, which they would try to 894 do anyway. 896 Other methods are possible. For example doing the selection of 897 Appointed Forwarders and of the TRILL switch in charge of that 898 selection across all TRILL switches on the link regardless of area. 899 However, a special case would then have to be made in any case for 900 legacy TRILL switches using area number zero. 902 Any of these techniques require multilevel aware RBridges to take 903 actions based on Hellos from from RBridges in other areas even though 904 they will not form an adjacency with such RBridges. 906 7. Summary 908 This draft discusses issues and possible approaches to multilevel 909 TRILL. The alternative using aggregated areas has significant 910 advantages in terms of scalability over using campus wide unique 911 nicknames, not just of avoiding nickname exhaustion, but by allowing 912 RPF Checks to be aggregated based on an entire area; however, the 913 alternative using unique nicknames is simpler and avoids the changes 914 in border TRILL switches required to support aggregated nicknames. 915 It is possible to support both. For example, a TRILL campus could use 916 simpler unique nicknames until scaling begins to cause problems and 917 then start to introduce areas with aggregated nicknames. 919 Some issues are not difficult, such as dealing with partitioned 920 areas. Some issues are more difficult, especially dealing with old 921 TRILL switches. 923 8. Security Considerations 925 This informational document explores alternatives for the use of 926 multilevel IS-IS in TRILL. It does not consider security issues. For 927 general TRILL Security Considerations, see [RFC6325]. 929 9. IANA Considerations 931 This document requires no IANA actions. 933 Normative References 935 [IS-IS] - ISO/IEC 10589:2002, Second Edition, "Intermediate System to 936 Intermediate System Intra-Domain Routing Exchange Protocol for 937 use in Conjunction with the Protocol for Providing the 938 Connectionless-mode Network Service (ISO 8473)", 2002. 940 [RFC6325] - Perlman, R., Eastlake 3rd, D., Dutt, D., Gai, S., and A. 941 Ghanwani, "Routing Bridges (RBridges): Base Protocol 942 Specification", RFC 6325, July 2011. 944 [RFC6439] - Perlman, R., Eastlake, D., Li, Y., Banerjee, A., and F. 945 Hu, "Routing Bridges (RBridges): Appointed Forwarders", RFC 946 6439, November 2011. 948 [rfc7180bis] - D. Eastlake, M. Zhang, et al, "TRILL: Clarifications, 949 Corrections, and Updates", draft-ietf-trill-rfc7180bis, work in 950 progress 952 Informative References 954 [RFC6361] - Carlson, J. and D. Eastlake 3rd, "PPP Transparent 955 Interconnection of Lots of Links (TRILL) Protocol Control 956 Protocol", RFC 6361, August 2011. 958 [RFC7172] - Eastlake 3rd, D., Zhang, M., Agarwal, P., Perlman, R., 959 and D. Dutt, "Transparent Interconnection of Lots of Links 960 (TRILL): Fine-Grained Labeling", RFC 7172, May 2014 962 [RFC7176] - Eastlake 3rd, D., Senevirathne, T., Ghanwani, A., Dutt, 963 D., and A. Banerjee, "Transparent Interconnection of Lots of 964 Links (TRILL) Use of IS-IS", RFC 7176, May 2014. 966 [RFC7177] - Eastlake 3rd, D., Perlman, R., Ghanwani, A., Yang, H., 967 and V. Manral, "Transparent Interconnection of Lots of Links 968 (TRILL): Adjacency", RFC 7177, May 2014, . 971 [RFC7357] - Zhai, H., Hu, F., Perlman, R., Eastlake 3rd, D., and O. 972 Stokes, "Transparent Interconnection of Lots of Links (TRILL): 973 End Station Address Distribution Information (ESADI) Protocol", 974 RFC 7357, September 2014, . 977 [DraftAggregated] - Bhargav Bhikkaji, Balaji Venkat Venkataswami, 978 Narayana Perumal Swamy, "Connecting Disparate Data 979 Center/PBB/Campus TRILL sites using BGP", draft-balaji-trill- 980 over-ip-multi-level, Work In Progress. 982 [DraftCMT] - Tissa Senevirathne, Janardhanan Pathang, Jon Hudson, 983 "Coordinated Multicast Trees (CMT) for TRILL", draft-tissa- 984 trill-cmt, Work in Progress. 986 [DraftUnique] - Tissa Senevirathne, Les Ginsberg, Janardhanan 987 Pathangi, Jon Hudson, Sam Aldrin, Ayan Banerjee, Sameer 988 Merchant, "Default Nickname Based Approach for Multilevel 989 TRILL", draft-tissa-trill-multilevel, Work In Progress. 991 [PseudoNickname] - H. Zhai, T. Senevirathne, et al, "TRILL: Pseudo- 992 Nickname for Active-active Access", draft-ietf-trill- 993 pseudonode-nickname, work in progress. 995 [NickFlags] - Eastlake, D., W. Hao, draft-eastlake-trill-nick-label- 996 prop, Work In Progress. 998 Acknowledgements 1000 The helpful comments of the following are hereby acknowledged: David 1001 Michael Bond and Dino Farinacci. 1003 The document was prepared in raw nroff. All macros used were defined 1004 within the source file. 1006 Authors' Addresses 1008 Radia Perlman 1009 EMC 1010 2010 256th Avenue NE, #200 1011 Bellevue, WA 98007 USA 1013 EMail: radia@alum.mit.edu 1015 Donald Eastlake 1016 Huawei Technologies 1017 155 Beaver Street 1018 Milford, MA 01757 USA 1020 Phone: +1-508-333-2270 1021 Email: d3e3e3@gmail.com 1023 Mingui Zhang 1024 Huawei Technologies 1025 No.156 Beiqing Rd. Haidian District, 1026 Beijing 100095 P.R. China 1028 EMail: zhangmingui@huawei.com 1030 Anoop Ghanwani 1031 Dell 1032 5450 Great America Parkway 1033 Santa Clara, CA 95054 USA 1035 EMail: anoop@alumni.duke.edu 1037 Hongjun Zhai 1038 ZTE 1039 68 Zijinghua Road, Yuhuatai District 1040 Nanjing, Jiangsu 210012 China 1042 Phone: +86 25 52877345 1043 Email: zhai.hongjun@zte.com.cn 1045 Copyright and IPR Provisions 1047 Copyright (c) 2015 IETF Trust and the persons identified as the 1048 document authors. All rights reserved. 1050 This document is subject to BCP 78 and the IETF Trust's Legal 1051 Provisions Relating to IETF Documents 1052 (http://trustee.ietf.org/license-info) in effect on the date of 1053 publication of this document. Please review these documents 1054 carefully, as they describe your rights and restrictions with respect 1055 to this document. Code Components extracted from this document must 1056 include Simplified BSD License text as described in Section 4.e of 1057 the Trust Legal Provisions and are provided without warranty as 1058 described in the Simplified BSD License. The definitive version of 1059 an IETF Document is that published by, or under the auspices of, the 1060 IETF. Versions of IETF Documents that are published by third parties, 1061 including those that are translated into other languages, should not 1062 be considered to be definitive versions of IETF Documents. The 1063 definitive version of these Legal Provisions is that published by, or 1064 under the auspices of, the IETF. Versions of these Legal Provisions 1065 that are published by third parties, including those that are 1066 translated into other languages, should not be considered to be 1067 definitive versions of these Legal Provisions. For the avoidance of 1068 doubt, each Contributor to the IETF Standards Process licenses each 1069 Contribution that he or she makes as part of the IETF Standards 1070 Process to the IETF Trust pursuant to the provisions of RFC 5378. No 1071 language to the contrary, or terms, conditions or rights that differ 1072 from or are inconsistent with the rights and licenses granted under 1073 RFC 5378, shall have any effect and shall be null and void, whether 1074 published or posted by such Contributor, or included with or in such 1075 Contribution.