idnits 2.17.1 draft-ietf-trill-directory-assisted-encap-08.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 : ---------------------------------------------------------------------------- No issues found here. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (January 17, 2018) is 2290 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) No issues found here. Summary: 0 errors (**), 0 flaws (~~), 1 warning (==), 1 comment (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 1 INTERNET-DRAFT Linda Dunbar 2 Intended status: Proposed Standard Donald Eastlake 3 Huawei 4 Radia Perlman 5 Dell/EMC 6 Expires: July 16, 2018 January 17, 2018 8 Directory Assisted TRILL Encapsulation 9 11 Abstract 13 This draft describes how data center networks can benefit from non- 14 RBridge nodes performing TRILL encapsulation with assistance from a 15 directory service. 17 Status of This Memo 19 This Internet-Draft is submitted to IETF in full conformance with the 20 provisions of BCP 78 and BCP 79. 22 Distribution of this document is unlimited. Comments should be sent 23 to the authors or the TRILL working group mailing list: 24 trill@ietf.org 26 Internet-Drafts are working documents of the Internet Engineering 27 Task Force (IETF), its areas, and its working groups. Note that 28 other groups may also distribute working documents as Internet- 29 Drafts. 31 Internet-Drafts are draft documents valid for a maximum of six months 32 and may be updated, replaced, or obsoleted by other documents at any 33 time. It is inappropriate to use Internet-Drafts as reference 34 material or to cite them other than as "work in progress." 36 The list of current Internet-Drafts can be accessed at 37 http://www.ietf.org/1id-abstracts.html. The list of Internet-Draft 38 Shadow Directories can be accessed at 39 http://www.ietf.org/shadow.html. 41 Table of Contents 43 1. Introduction............................................3 44 2. Conventions Used in This Document.......................4 46 3. Directory Assistance to Non-RBridge.....................5 47 4. Source Nickname in Encapsulation by Non-RBridge Nodes...8 49 5. Benefits of Non-RBridge Performing TRILL Encapsulation..9 50 5.1. Avoid Nickname Exhaustion Issue.......................9 51 5.2. Reduce MAC Tables for Switches on Bridged LANs........9 53 6. Manageability Considerations...........................11 54 7. Security Considerations................................11 55 8. IANA Considerations....................................12 57 Normative References......................................13 58 Informative References....................................13 60 Acknowledgments...........................................13 61 Authors' Addresses........................................14 63 1. Introduction 65 This document describes how data center networks can benefit from 66 non-RBridge nodes performing TRILL encapsulation with assistance from 67 a directory service and specifies a method for them to do so. 69 [RFC7067] and [RFC8171] describe the framework and methods for edge 70 RBridges to get MAC&VLAN <-> Edge RBridge mapping from a directory 71 service instead of flooding unknown destination MAC addresses across 72 a TRILL domain. If it has the needed directory information, any node, 73 even a non-RBridge node, can perform the TRILL data packet 74 encapsulation. This draft describes the benefits of and a scheme for 75 non-RBridge nodes performing TRILL encapsulation. 77 2. Conventions Used in This Document 79 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 80 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 81 document are to be interpreted as described in [RFC2119]. 83 AF: Appointed Forwarder RBridge port [RFC8139]. 85 Bridge: An IEEE 802.1Q compliant device. In this draft, Bridge is 86 used interchangeably with Layer 2 switch. 88 DA: Destination Address. 90 ES-IS: End System to Intermediate Systems [RFC8171]. 92 Host: A physical server or a virtual machine running 93 applications. A host usually has at least one IP address 94 and at least one MAC address. 96 IS-IS: Intermediate System to Intermediate System [RFC7176]. 98 SA: Source Address. 100 TRILL-EN: TRILL Encapsulating node. A node that performs the TRILL 101 encapsulation but doesn't participate in RBridge's IS-IS 102 routing. 104 VM: Virtual Machine. 106 3. Directory Assistance to Non-RBridge 108 With directory assistance [RFC7067] [RFC8171], a non-RBridge node can 109 learn if a data packet needs to be forwarded across the RBridge 110 domain and if so the corresponding egress RBridge. 112 Suppose the RBridge domain boundary starts at network switches (not 113 virtual switches embedded on servers). (See Figure 1 for a high level 114 diagram of a typical data center network.) A directory can assist 115 Virtual Switches embedded on servers to encapsulate with a proper 116 TRILL header by providing the nickname of the egress RBridge edge to 117 which the destination is attached. The other information needed to 118 encapsulate can be either learned by listening to TRILL ES-IS and/or 119 IS-IS Hellos [RFC7176] [RFC8171], which will indicate the MAC address 120 and nickname of appropriate local edge RBridges, or by configuration. 122 If it is not known whether a destination is attached to one or more 123 RBridge edge nodes, based on the directory, the non-RBridge node can 124 forward the data frames natively, i.e. not encapsulating with any 125 TRILL header. Or, if the directory is known to be complete, the non- 126 RBridge node can discard such data frames. 128 \ +-------+ +------+ TRILL Domain/ 129 \ +/------+ | +/-----+ | / 130 \ | Aggr11| + ----- |AggrN1| + / 131 \ +---+---+/ +------+/ / 132 \ / \ / \ / 133 \ / \ / \ / 134 \ +---+ +---+ +---+ +---+ / 135 \- |T11|... |T1x| |T21| .. |T2y|--- 136 +---+ +---+ +---+ +---+ 137 | | | | 138 +-|-+ +-|-+ +-|-+ +-|-+ 139 | |... | V | | V | .. | V |<- vSwitch 140 +---+ +---+ +---+ +---+ 141 | |... | V | | V | .. | V | 142 +---+ +---+ +---+ +---+ 143 | |... | V | | V | .. | V | 144 +---+ +---+ +---+ +---+ 146 Figure 1. TRILL domain in a typical Data Center Network 148 When a TRILL encapsulated data packet reaches the ingress RBridge, 149 that RBridge simply performs the usual TRILL processing and forwards 150 the pre-encapsulated packet to the RBridge that is specified by the 151 egress nickname field of the TRILL header. When an ingress RBridge 152 receives a native Ethernet frame in an environment with complete 153 directory information, the ingress RBridge doesn't flood or forward 154 the received data frames when the destination MAC address in the 155 Ethernet data frames is unknown. 157 When all end nodes attached to an ingress RBridge pre-encapsulate 158 with a TRILL header for traffic across the TRILL domain, the ingress 159 RBridge doesn't need to encapsulate any native Ethernet frames to the 160 TRILL domain. The attached nodes can be connected to multiple edge 161 RBridges by having multiple ports or through a bridged LAN. All 162 RBridge edge ports connected to one bridged LAN can receive and 163 forward pre-encapsulated traffic, which can greatly improve the 164 overall network utilization. However, it is still necessary to 165 designate AF ports to, for example, be sure that multi-destination 166 packets from the TRILL campus are only egressed through one RBridge. 168 The TRILL base protocol specification [RFC6325] Section 4.6.2 Bullet 169 8 specifies that an RBridge port can be configured to accept TRILL 170 encapsulated frames from a neighbor that is not an RBridge. 172 When a TRILL frame arrives at an RBridge whose nickname matches the 173 destination nickname in the TRILL header of the frame, the processing 174 is exactly as normal: as specified in [RFC6325] the RBridge 175 decapsulates the received TRILL frame and forwards the decapsulated 176 frame to the target attached to its edge ports. When the destination 177 MAC address of the decapsulated Ethernet frame is not in the egress 178 RBridge's local MAC attachment tables, the egress RBridge floods the 179 decapsulated frame to all attached links in the frame's VLAN, or 180 drops the frame (if the egress RBridge is configured with that 181 policy). 183 We call a node that, as specified herein, only performs TRILL 184 encapsulation, but doesn't participate in RBridge's IS-IS routing, a 185 TRILL Encapsulating node (TRILL-EN). The TRILL Encapsulating Node can 186 pull MAC&VLAN <-> Edge RBridge mapping from directory servers 187 [RFC8171]. In order to do this, a TRILL-EN MUST support TRILL ES-IS 188 [RFC8171]. 190 Upon receiving or locally generating a native Ethernet frame, the 191 TRILL-EN checks the MAC&VLAN <-> Edge RBridge mapping, and performs 192 the corresponding TRILL encapsulation if the mapping entry is found 193 as snown in Figure 2. If the destination MAC address and VLAN of the 194 received Ethernet frame doesn't exist in the mapping table and there 195 is no positive reply from pull requests to a directory, the Ethernet 196 frame is dropped or is forwarded in native form to an edge RBridge, 197 depending on the TRILL-EN configuration. 199 +------------+--------+---------+---------+--+-------+---+ 200 |OuterEtherHd|TRILL HD| InnerDA | InnerSA |..|Payload|FCS| 201 +------------+--------+---------+---------+--+-------+---+ 202 | 203 | | | 204 | 205 | 206 | +-------+ TRILL +------+ 207 | | R1 |-----------| R2 | Decapsulate 208 | +---+---+ domain +------+ TRILL header 209 v | | 210 +---------->| | 211 | | 212 +-----+ +-----+ 213 Non-RBridge node: |T12 | | T22 | 214 Encapsulate TRILL +-----+ +-----+ 215 Header for data 216 Frames to traverse TRILL domain. 218 Figure 2. Data frames from a TRILL-EN 220 4. Source Nickname in Encapsulation by Non-RBridge Nodes 222 The TRILL header includes a Source RBridge's Nickname (ingress) and 223 Destination RBridge's Nickname (egress). When a TRILL header is added 224 to a data packet by a TRILL-EN, the Ingress RBridge nickname field in 225 the TRILL header is set to a nickname of the AF for the data packet's 226 VLAN. The TRILL-EN determines the AF by snooping on IS-IS Hellos from 227 the edge RBridges on the link with the TRILL-EN in the same way that 228 the RBridges on the link determine the AF [RFC8139]. A TRILL-EN is 229 free to send the encapsulated data frame to any of the edge RBridges 230 on its link. 232 5. Benefits of Non-RBridge Performing TRILL Encapsulation 234 This section summarizing benefits of having a non-RBridge node 235 perform TRILL encapsulation. 237 5.1. Avoid Nickname Exhaustion Issue 239 For a large Data Center with hundreds of thousands of virtualized 240 servers, setting the TRILL boundary at the servers' virtual switches 241 will create a TRILL domain with hundreds of thousands of RBridge 242 nodes, which has issues of TRILL Nickname exhaustion and challenges 243 to IS-IS. On the other hand, setting the TRILL boundary at 244 aggregation switches that have many virtualized servers attached can 245 limit the number of RBridge nodes in a TRILL domain, but introduces 246 the issue of very large MAC&VLAN <-> Edge RBridge mapping tables to 247 be maintained by RBridge edge nodes. 249 Allowing Non-RBridge nodes to pre-encapsulate data frames with TRILL 250 headers makes it possible to have a TRILL domain with a reasonable 251 number of RBridge nodes in a large data center. All the TRILL-ENs 252 attached to one RBridge can be represented by one TRILL nickname, 253 which can avoid the Nickname exhaustion problem. 255 5.2. Reduce MAC Tables for Switches on Bridged LANs 257 When hosts in a VLAN (or subnet) span across multiple RBridge edge 258 nodes and each RBridge edge has multiple VLANs enabled, the switches 259 on the bridged LANs attached to the RBridge edge are exposed to all 260 MAC addresses among all the VLANs enabled. 262 For example, for an Access Switch with 40 physical servers attached, 263 where each server has 100 VMs, there are 4000 hosts under the Access 264 Switch. If indeed hosts/VMs can be moved anywhere, the worst case for 265 the Access Switch is when all those 4000 VMs belong to different 266 VLANs, i.e. the access switch has 4000 VLANs enabled. If each VLAN 267 has 200 hosts, this access switch's MAC table potentially has 268 200*4000 = 800,000 entries. 270 If the virtual switches on servers pre-encapsulate the data frames 271 destined for hosts attached to remote RBridge Edge nodes, the outer 272 MAC destination address of those TRILL encapsulated data frames will 273 be the MAC address of a local RBridge edge, i.e. the ingress 274 RBridge. The switches on the local bridged LAN don't need to keep the 275 MAC entries for remote hosts attached to other edge RBridges. 277 But the TRILL traffic from nodes attached to other RBridges is 278 decapsulated and has the true source and destination MACs. One simple 279 way to prevent local bridges from learning remote hosts' MACs and 280 adding to their MAC tables, if that would be a problem, is to disable 281 this data plane learning on local bridges. The local bridges can be 282 pre-configured with MAC addresses of local hosts with the assistance 283 of a directory. The local bridges can always send frames with 284 unknown destination MAC addresses to the ingress RBridge. In an 285 environment where a large number of VMs are instantiated in one 286 server, the number of remote MAC addresses could be very large. If it 287 is not feasible to disable learning and pre- configure MAC tables for 288 local bridges and all important traffic is IP, one effective method 289 to minimize local bridges' MAC table size is to use the server's MAC 290 address to hide MAC addresses of the attached VMs. I.e., the server 291 acting as an edge node uses its own MAC address in the source MAC 292 address field of the packets originated from a host (or VM) embedded. 293 When the Ethernet frame arrives at the target edge node (the egress), 294 the target edge node can send the packet to the corresponding 295 destination host based on the packet's IP address. Very often, the 296 target edge node communicates with the embedded VMs via a layer 2 297 virtual switch. In this case, the target edge node can construct the 298 proper Ethernet header with the assistance of the directory. The 299 information from the directory includes the proper host IP to MAC 300 mapping information. 302 6. Manageability Considerations 304 Directory assistance [RFC8171] is required to make it possible for a 305 non-TRILL node to pre-encapsulate packets destined towards remote 306 RBridges. TRILL-ENs have the same configuration options as any pull 307 directory client. See Section 4 of [RFC8171]. 309 7. Security Considerations 311 The mechanism described in this document requires TRILL-ENs to be 312 aware of the MAC address(es) of the TRILL edge RBridge(s) to which 313 the TRILL-EN is attached and the egress RBridge nickname from which 314 the destination of the packets is reachable. With that information, 315 TRILL-ENs can learn a substantial amount about the topology of the 316 TRILL domain. Therefore, there could be a potential security risk 317 when the TRILL-ENs are not trusted. In addition, if the path between 318 the directory and the TRILL-ENs are attacked, false mappings can be 319 sent to the TRILL-EN causing packets from the TRILL-EN to be sent to 320 wrong destinations, possibly violating security policy. Therefore, a 321 combination of authentication and encryption should be used between 322 the Directory and TRILL-EN. The entities involved will need to 323 properly authenticate with each other to protect sensitive 324 information. 326 Use of directory assisted ecapslation by TRILL-ENs essentially 327 involves those TRILL-ENs spoofing edge RBridges to which they are 328 connected, which is another reason that TRILL-ENs should be trusted 329 nodes. Such spoofing cannot cause looping traffic because TRILL has a 330 hop count in the TRILL header [RFC6325] so that, should there be a 331 loop, a TRILL packet caught in that loop (i.e., an encapsulated 332 frame) will be discarded. (In the potentially more dangerous case of 333 multi-destination packets, as compared with known unicast, where 334 copies could multiply due to forks in the distribution tree, a 335 Reverse Path Forwarding Check is also used [RFC6325] to discard 336 packets that appear to be on the wrong link or when there is 337 disagreement about the distribution tree.) 339 For Pull Directory and TRILL ES-IS security considerations, see 340 [RFC8171]. 342 For general TRILL security considerations, see [RFC6325]. 344 8. IANA Considerations 346 This document requires no IANA actions. RFC Editor: please remove 347 this section before publication. 349 Normative References 351 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 352 Requirement Levels", BCP 14, RFC 2119, DOI 10.17487/RFC2119, 353 March 1997, . 355 [RFC6325] Perlman, R., Eastlake 3rd, D., Dutt, D., Gai, S., and A. 356 Ghanwani, "Routing Bridges (RBridges): Base Protocol 357 Specification", RFC 6325, DOI 10.17487/RFC6325, July 2011, 358 . 360 [RFC7176] Eastlake 3rd, D., Senevirathne, T., Ghanwani, A., Dutt, D., 361 and A. Banerjee, "Transparent Interconnection of Lots of Links 362 (TRILL) Use of IS-IS", RFC 7176, DOI 10.17487/RFC7176, May 363 2014, . 365 [RFC8139] Eastlake 3rd, D., Li, Y., Umair, M., Banerjee, A., and F. 366 Hu, "Transparent Interconnection of Lots of Links (TRILL): 367 Appointed Forwarders", RFC 8139, DOI 10.17487/RFC8139, June 368 2017, . 370 [RFC8171] Eastlake 3rd, D., Dunbar, L., Perlman, R., and Y. Li, 371 "Transparent Interconnection of Lots of Links (TRILL): Edge 372 Directory Assistance Mechanisms", RFC 8171, DOI 373 10.17487/RFC8171, June 2017, . 376 Informative References 378 [RFC7067] Dunbar, L., Eastlake 3rd, D., Perlman, R., and I. 379 Gashinsky, "Directory Assistance Problem and High-Level Design 380 Proposal", RFC 7067, DOI 10.17487/RFC7067, November 2013, 381 . 383 Acknowledgments 385 The following are thanked for their contributions: 387 Igor Gashinsky 388 Ben Nevin-Jenkins 390 The document was prepared in raw nroff. All macros used were defined 391 within the source file. 393 Authors' Addresses 395 Linda Dunbar 396 Huawei Technologies 397 5340 Legacy Drive, Suite 175 398 Plano, TX 75024, USA 400 Phone: +1-469-277-5840 401 Email: linda.dunbar@huawei.com 403 Donald Eastlake 404 Huawei Technologies 405 155 Beaver Street 406 Milford, MA 01757 USA 408 Phone: +1-508-333-2270 409 Email: d3e3e3@gmail.com 411 Radia Perlman 412 Dell/EMC 413 2010 256th Avenue NE, #200 414 Bellevue, WA 98007 USA 416 Email: Radia@alum.mit.edu 418 Copyright, Disclaimer, and Additional IPR Provisions 420 Copyright (c) 2018 IETF Trust and the persons identified as the 421 document authors. All rights reserved. 423 This document is subject to BCP 78 and the IETF Trust's Legal 424 Provisions Relating to IETF Documents 425 (http://trustee.ietf.org/license-info) in effect on the date of 426 publication of this document. Please review these documents 427 carefully, as they describe your rights and restrictions with respect 428 to this document. Code Components extracted from this document must 429 include Simplified BSD License text as described in Section 4.e of 430 the Trust Legal Provisions and are provided without warranty as 431 described in the Simplified BSD License. The definitive version of 432 an IETF Document is that published by, or under the auspices of, the 433 IETF. Versions of IETF Documents that are published by third parties, 434 including those that are translated into other languages, should not 435 be considered to be definitive versions of IETF Documents. The 436 definitive version of these Legal Provisions is that published by, or 437 under the auspices of, the IETF. Versions of these Legal Provisions 438 that are published by third parties, including those that are 439 translated into other languages, should not be considered to be 440 definitive versions of these Legal Provisions. For the avoidance of 441 doubt, each Contributor to the IETF Standards Process licenses each 442 Contribution that he or she makes as part of the IETF Standards 443 Process to the IETF Trust pursuant to the provisions of RFC 5378. No 444 language to the contrary, or terms, conditions or rights that differ 445 from or are inconsistent with the rights and licenses granted under 446 RFC 5378, shall have any effect and shall be null and void, whether 447 published or posted by such Contributor, or included with or in such 448 Contribution.