idnits 2.17.1 draft-ietf-forces-interfelfb-06.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-RFC2606-compliant FQDNs in the document. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year == Line 854 has weird spacing: '...putPort group...' -- The document date (July 1, 2016) is 2856 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: 'RFC 791' is mentioned on line 463, but not defined -- Possible downref: Non-RFC (?) normative reference: ref. 'ieee8021ae' -- Possible downref: Non-RFC (?) normative reference: ref. 'ieee8021x' -- Obsolete informational reference (is this intentional?): RFC 2460 (Obsoleted by RFC 8200) Summary: 0 errors (**), 0 flaws (~~), 4 warnings (==), 4 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Internet Engineering Task Force D. Joachimpillai 3 Internet-Draft Verizon 4 Intended status: Standards Track J. Hadi Salim 5 Expires: January 2, 2017 Mojatatu Networks 6 July 1, 2016 8 ForCES Inter-FE LFB 9 draft-ietf-forces-interfelfb-06 11 Abstract 13 This document describes how to extend the ForCES LFB topology across 14 FEs by defining the Inter-FE LFB Class. The Inter-FE LFB Class 15 provides the ability to pass data and metadata across FEs without 16 needing any changes to the ForCES specification. The document 17 focuses on Ethernet transport. 19 Status of This Memo 21 This Internet-Draft is submitted in full conformance with the 22 provisions of BCP 78 and BCP 79. 24 Internet-Drafts are working documents of the Internet Engineering 25 Task Force (IETF). Note that other groups may also distribute 26 working documents as Internet-Drafts. The list of current Internet- 27 Drafts is at http://datatracker.ietf.org/drafts/current/. 29 Internet-Drafts are draft documents valid for a maximum of six months 30 and may be updated, replaced, or obsoleted by other documents at any 31 time. It is inappropriate to use Internet-Drafts as reference 32 material or to cite them other than as "work in progress." 34 This Internet-Draft will expire on January 2, 2017. 36 Copyright Notice 38 Copyright (c) 2016 IETF Trust and the persons identified as the 39 document authors. All rights reserved. 41 This document is subject to BCP 78 and the IETF Trust's Legal 42 Provisions Relating to IETF Documents 43 (http://trustee.ietf.org/license-info) in effect on the date of 44 publication of this document. Please review these documents 45 carefully, as they describe your rights and restrictions with respect 46 to this document. Code Components extracted from this document must 47 include Simplified BSD License text as described in Section 4.e of 48 the Trust Legal Provisions and are provided without warranty as 49 described in the Simplified BSD License. 51 Table of Contents 53 1. Terminology and Conventions . . . . . . . . . . . . . . . . . 2 54 1.1. Requirements Language . . . . . . . . . . . . . . . . . . 3 55 1.2. Definitions . . . . . . . . . . . . . . . . . . . . . . . 3 56 2. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 3 57 3. Problem Scope And Use Cases . . . . . . . . . . . . . . . . . 4 58 3.1. Assumptions . . . . . . . . . . . . . . . . . . . . . . . 4 59 3.2. Sample Use Cases . . . . . . . . . . . . . . . . . . . . 4 60 3.2.1. Basic IPv4 Router . . . . . . . . . . . . . . . . . . 4 61 3.2.1.1. Distributing The Basic IPv4 Router . . . . . . . 6 62 3.2.2. Arbitrary Network Function . . . . . . . . . . . . . 7 63 3.2.2.1. Distributing The Arbitrary Network Function . . . 8 64 4. Inter-FE LFB Overview . . . . . . . . . . . . . . . . . . . . 8 65 4.1. Inserting The Inter-FE LFB . . . . . . . . . . . . . . . 9 66 5. Inter-FE Ethernet Connectivity . . . . . . . . . . . . . . . 10 67 5.1. Inter-FE Ethernet Connectivity Issues . . . . . . . . . . 10 68 5.1.1. MTU Consideration . . . . . . . . . . . . . . . . . . 11 69 5.1.2. Quality Of Service Considerations . . . . . . . . . . 11 70 5.1.3. Congestion Considerations . . . . . . . . . . . . . . 11 71 5.2. Inter-FE Ethernet Encapsulation . . . . . . . . . . . . . 12 72 6. Detailed Description of the Ethernet inter-FE LFB . . . . . . 13 73 6.1. Data Handling . . . . . . . . . . . . . . . . . . . . . . 14 74 6.1.1. Egress Processing . . . . . . . . . . . . . . . . . . 14 75 6.1.2. Ingress Processing . . . . . . . . . . . . . . . . . 15 76 6.2. Components . . . . . . . . . . . . . . . . . . . . . . . 16 77 6.3. Inter-FE LFB XML Model . . . . . . . . . . . . . . . . . 17 78 7. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 21 79 8. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 21 80 9. IEEE Assignment Considerations . . . . . . . . . . . . . . . 22 81 10. Security Considerations . . . . . . . . . . . . . . . . . . . 22 82 11. References . . . . . . . . . . . . . . . . . . . . . . . . . 23 83 11.1. Normative References . . . . . . . . . . . . . . . . . . 23 84 11.2. Informative References . . . . . . . . . . . . . . . . . 24 85 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 25 87 1. Terminology and Conventions 88 1.1. Requirements Language 90 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 91 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 92 document are to be interpreted as described in [RFC2119]. 94 1.2. Definitions 96 This document depends on the terminology defined in several ForCES 97 documents [RFC3746], [RFC5810], [RFC5811], and [RFC5812] [RFC7391] 98 [RFC7408] for the sake of contextual clarity. 100 Control Engine (CE) 102 Forwarding Engine (FE) 104 FE Model 106 LFB (Logical Functional Block) Class (or type) 108 LFB Instance 110 LFB Model 112 LFB Metadata 114 ForCES Component 116 LFB Component 118 ForCES Protocol Layer (ForCES PL) 120 ForCES Protocol Transport Mapping Layer (ForCES TML) 122 2. Introduction 124 In the ForCES architecture, a packet service can be modelled by 125 composing a graph of one or more LFB instances. The reader is 126 referred to the details in the ForCES Model [RFC5812]. 128 The ForCES model describes the processing within a single Forwarding 129 Element (FE) in terms of logical forwarding blocks (LFB), including 130 provision for the Control Element (CE) to establish and modify that 131 processing sequence, and the parameters of the individual LFBs. 133 Under some circumstance, it would be beneficial to be able to extend 134 this view, and the resulting processing across more than one FE. 135 This may be in order to achieve scale by splitting the processing 136 across elements, or to utilize specialized hardware available on 137 specific FEs. 139 Given that the ForCES inter-LFB architecture calls for the ability to 140 pass metadata between LFBs, it is imperative therefore to define 141 mechanisms to extend that existing feature and allow passing the 142 metadata between LFBs across FEs. 144 This document describes how to extend the LFB topology across FEs i.e 145 inter-FE connectivity without needing any changes to the ForCES 146 definitions. It focuses on using Ethernet as the interconnection 147 between FEs. 149 3. Problem Scope And Use Cases 151 The scope of this document is to solve the challenge of passing 152 ForCES defined metadata alongside packet data across FEs (be they 153 physical or virtual) for the purpose of distributing the LFB 154 processing. 156 3.1. Assumptions 158 o The FEs involved in the Inter-FE LFB belong to the same Network 159 Element(NE) and are within a single administrative private network 160 which is in close proximity. 162 o The FEs are already interconnected using Ethernet. We focus on 163 Ethernet because it is a very common setup as an FE interconnect. 164 Other higher transports (such as UDP over IP) or lower transports 165 could be defined to carry the data and metadata, but these cases 166 are not addressed in this document. 168 3.2. Sample Use Cases 170 To illustrate the problem scope we present two use cases where we 171 start with a single FE running all the LFBs functionality then split 172 it into multiple FEs achieving the same end goals. 174 3.2.1. Basic IPv4 Router 176 A sample LFB topology depicted in Figure 1 demonstrates a service 177 graph for delivering basic IPv4 forwarding service within one FE. 178 For the purpose of illustration, the diagram shows LFB classes as 179 graph nodes instead of multiple LFB class instances. 181 Since the illustration on Figure 1 is meant only as an exercise to 182 showcase how data and metadata are sent down or upstream on a graph 183 of LFB instances, it abstracts out any ports in both directions and 184 talks about a generic ingress and egress LFB. Again, for 185 illustration purposes, the diagram does not show exception or error 186 paths. Also left out are details on Reverse Path Filtering, ECMP, 187 multicast handling etc. In other words, this is not meant to be a 188 complete description of an IPv4 forwarding application; for a more 189 complete example, please refer the LFBlib document [RFC6956]. 191 The output of the ingress LFB(s) coming into the IPv4 Validator LFB 192 will have both the IPv4 packets and, depending on the implementation, 193 a variety of ingress metadata such as offsets into the different 194 headers, any classification metadata, physical and virtual ports 195 encountered, tunnelling information etc. These metadata are lumped 196 together as "ingress metadata". 198 Once the IPv4 validator vets the packet (example ensures that no 199 expired TTL etc), it feeds the packet and inherited metadata into the 200 IPv4 unicast LPM LFB. 202 +----+ 203 | | 204 IPv4 pkt | | IPv4 pkt +-----+ +---+ 205 +------------->| +------------->| | | | 206 | + ingress | | + ingress |IPv4 | IPv4 pkt | | 207 | metadata | | metadata |Ucast+------------>| +--+ 208 | +----+ |LPM | + ingress | | | 209 +-+-+ IPv4 +-----+ + NHinfo +---+ | 210 | | Validator metadata IPv4 | 211 | | LFB NextHop| 212 | | LFB | 213 | | | 214 | | IPv4 pkt | 215 | | + {ingress | 216 +---+ + NHdetails} 217 Ingress metadata | 218 LFB +--------+ | 219 | Egress | | 220 <--+ |<-----------------+ 221 | LFB | 222 +--------+ 224 Figure 1: Basic IPv4 packet service LFB topology 226 The IPv4 unicast LPM LFB does a longest prefix match lookup on the 227 IPv4 FIB using the destination IP address as a search key. The 228 result is typically a next hop selector which is passed downstream as 229 metadata. 231 The Nexthop LFB receives the IPv4 packet with an associated next hop 232 info metadata. The NextHop LFB consumes the NH info metadata and 233 derives from it a table index to look up the next hop table in order 234 to find the appropriate egress information. The lookup result is 235 used to build the next hop details to be used downstream on the 236 egress. This information may include any source and destination 237 information (for our purposes, MAC addresses to use) as well as 238 egress ports. [Note: It is also at this LFB where typically the 239 forwarding TTL decrementing and IP checksum recalculation occurs.] 241 The details of the egress LFB are considered out of scope for this 242 discussion. Suffice it is to say that somewhere within or beyond the 243 Egress LFB the IPv4 packet will be sent out a port (Ethernet, virtual 244 or physical etc). 246 3.2.1.1. Distributing The Basic IPv4 Router 248 Figure 2 demonstrates one way the router LFB topology in Figure 1 may 249 be split across two FEs (eg two ASICs). Figure 2 shows the LFB 250 topology split across FEs after the IPv4 unicast LPM LFB. 252 FE1 253 +-------------------------------------------------------------+ 254 | +----+ | 255 | +----------+ | | | 256 | | Ingress | IPv4 pkt | | IPv4 pkt +-----+ | 257 | | LFB +-------------->| +------------->| | | 258 | | | + ingress | | + ingress |IPv4 | | 259 | +----------+ metadata | | metadata |Ucast| | 260 | ^ +----+ |LPM | | 261 | | IPv4 +--+--+ | 262 | | Validator | | 263 | LFB | | 264 +---------------------------------------------------|---------+ 265 | 266 IPv4 packet + 267 {ingress + NHinfo} 268 metadata 269 FE2 | 270 +---------------------------------------------------|---------+ 271 | V | 272 | +--------+ +--------+ | 273 | | Egress | IPv4 packet | IPv4 | | 274 | <-----+ LFB |<----------------------+NextHop | | 275 | | |{ingress + NHdetails} | LFB | | 276 | +--------+ metadata +--------+ | 277 +-------------------------------------------------------------+ 279 Figure 2: Split IPv4 packet service LFB topology 281 Some proprietary inter-connect (example Broadcom HiGig over XAUI 282 [brcm-higig]) are known to exist to carry both the IPv4 packet and 283 the related metadata between the IPv4 Unicast LFB and IPv4 NextHop 284 LFB across the two FEs. 286 This document defines the inter-FE LFB, a standard mechanism for 287 encapsulating, generating, receiving and decapsulating packets and 288 associated metadata FEs over Ethernet. 290 3.2.2. Arbitrary Network Function 292 In this section we show an example of an arbitrary Network Function 293 which is more coarse grained in terms of functionality. Each Network 294 Function may constitute more than one LFB. 296 FE1 297 +-------------------------------------------------------------+ 298 | +----+ | 299 | +----------+ | | | 300 | | Network | pkt |NF2 | pkt +-----+ | 301 | | Function +-------------->| +------------->| | | 302 | | 1 | + NF1 | | + NF1/2 |NF3 | | 303 | +----------+ metadata | | metadata | | | 304 | ^ +----+ | | | 305 | | +--+--+ | 306 | | | | 307 | | | 308 +---------------------------------------------------|---------+ 309 V 311 Figure 3: A Network Function Service Chain within one FE 313 The setup in Figure 3 is a typical of most packet processing boxes 314 where we have functions like DPI, NAT, Routing, etc connected in such 315 a topology to deliver a packet processing service to flows. 317 3.2.2.1. Distributing The Arbitrary Network Function 319 The setup in Figure 3 can be split out across 3 FEs instead of as 320 demonstrated in Figure 4. This could be motivated by scale out 321 reasons or because different vendors provide different functionality 322 which is plugged-in to provide such functionality. The end result is 323 to have the same packet service delivered to the different flows 324 passing through. 326 FE1 FE2 327 +----------+ +----+ FE3 328 | Network | pkt |NF2 | pkt +-----+ 329 | Function +-------------->| +------------->| | 330 | 1 | + NF1 | | + NF1/2 |NF3 | 331 +----------+ metadata | | metadata | | 332 ^ +----+ | | 333 | +--+--+ 334 | 335 V 337 Figure 4: A Network Function Service Chain Distributed Across 338 Multiple FEs 340 4. Inter-FE LFB Overview 342 We address the inter-FE connectivity requirements by defining the 343 inter-FE LFB class. Using a standard LFB class definition implies no 344 change to the basic ForCES architecture in the form of the core LFBs 345 (FE Protocol or Object LFBs). This design choice was made after 346 considering an alternative approach that would have required changes 347 to both the FE Object capabilities (SupportedLFBs) as well 348 LFBTopology component to describe the inter-FE connectivity 349 capabilities as well as runtime topology of the LFB instances. 351 4.1. Inserting The Inter-FE LFB 353 The distributed LFB topology described in Figure 2 is re-illustrated 354 in Figure 5 to show the topology location where the inter-FE LFB 355 would fit in. 357 As can be observed in Figure 5, the same details passed between IPv4 358 unicast LPM LFB and the IPv4 NH LFB are passed to the egress side of 359 the Inter-FE LFB. This information is illustrated as multiplicity of 360 inputs into the egress InterFE LFB instance. Each input represents a 361 unique set of selection information. 363 FE1 364 +-------------------------------------------------------------+ 365 | +----------+ +----+ | 366 | | Ingress | IPv4 pkt | | IPv4 pkt +-----+ | 367 | | LFB +-------------->| +------------->| | | 368 | | | + ingress | | + ingress |IPv4 | | 369 | +----------+ metadata | | metadata |Ucast| | 370 | ^ +----+ |LPM | | 371 | | IPv4 +--+--+ | 372 | | Validator | | 373 | | LFB | | 374 | | IPv4 pkt + metadata | 375 | | {ingress + NHinfo} | 376 | | | | 377 | | +..--+..+ | 378 | | |..| | | | 379 | +-V--V-V--V-+ | 380 | | Egress | | 381 | | InterFE | | 382 | | LFB | | 383 | +------+----+ | 384 +---------------------------------------------------|---------+ 385 | 386 Ethernet Frame with: | 387 IPv4 packet data and metadata 388 {ingress + NHinfo + Inter FE info} 389 FE2 | 390 +---------------------------------------------------|---------+ 391 | +..+.+..+ | 392 | |..|.|..| | 393 | +-V--V-V--V-+ | 394 | | Ingress | | 395 | | InterFE | | 396 | | LFB | | 397 | +----+------+ | 398 | | | 399 | IPv4 pkt + metadata | 400 | {ingress + NHinfo} | 401 | | | 402 | +--------+ +----V---+ | 403 | | Egress | IPv4 packet | IPv4 | | 404 | <-----+ LFB |<----------------------+NextHop | | 405 | | |{ingress + NHdetails} | LFB | | 406 | +--------+ metadata +--------+ | 407 +-------------------------------------------------------------+ 409 Figure 5: Split IPv4 forwarding service with Inter-FE LFB 411 The egress of the inter-FE LFB uses the received packet and metadata 412 to select details for encapsulation when sending messages towards the 413 selected neighboring FE. These details include what to communicate 414 as the source and destination FEs (abstracted as MAC addresses as 415 described in Section 5.2); in addition the original metadata may be 416 passed along with the original IPv4 packet. 418 On the ingress side of the inter-FE LFB the received packet and its 419 associated metadata are used to decide the packet graph continuation. 420 This includes which of the original metadata and which next LFB class 421 instance to continue processing on. In the illustrated Figure 5, an 422 IPv4 Nexthop LFB instance is selected and appropriate metadata is 423 passed on to it. 425 The ingress side of the inter-FE LFB consumes some of the information 426 passed and passes on the IPv4 packet alongside with the ingress and 427 NHinfo metadata to the IPv4 NextHop LFB as was done earlier in both 428 Figure 1 and Figure 2. 430 5. Inter-FE Ethernet Connectivity 432 Section 5.1 describes some of the issues related to using Ethernet as 433 the transport and how we mitigate them. 435 Section 5.2 defines a payload format that is to be used over 436 Ethernet. An existing implementation of this specification on top of 437 Linux Traffic Control [linux-tc] is described in [tc-ife]. 439 5.1. Inter-FE Ethernet Connectivity Issues 441 There are several issues that may occur due to using direct Ethernet 442 encapsulation that need consideration. 444 5.1.1. MTU Consideration 446 Because we are adding data to existing Ethernet frames, MTU issues 447 may arise. We recommend: 449 o To use large MTUs when possible (example with jumbo frames). 451 o Limit the amount of metadata that could be transmitted; our 452 definition allows for filtering of select metadata to be 453 encapsulated in the frame as described in Section 6. We recommend 454 sizing the egress port MTU so as to allow space for maximum size 455 of the metadata total size to allow between FEs. In such a setup, 456 the port is configured to "lie" to the upper layers by claiming to 457 have a lower MTU than it is capable of. MTU setting can be 458 achieved by ForCES control of the port LFB(or other config). In 459 essence, the control plane when explicitly making a decision for 460 the MTU settings of the egress port is implicitly deciding how 461 much metadata will be allowed. Caution needs to be exercised on 462 how low the resulting reported link MTU could be: For IPv4 packets 463 the minimum size is 64 octets [RFC 791] and for IPv6 the minimum 464 size is 1280 octets [RFC2460]. 466 5.1.2. Quality Of Service Considerations 468 A raw packet arriving at the Inter-FE LFB (from upstream LFB Class 469 instances) may have COS metadatum indicating how it should be treated 470 from a Quality of Service perspective. 472 The resulting Ethernet frame will be eventually (preferentially) 473 treated by a downstream LFB(typically a port LFB instance) and their 474 COS marks will be honored in terms of priority. In other words the 475 presence of the Inter-FE LFB does not change the COS semantics 477 5.1.3. Congestion Considerations 479 Most of the traffic passing through FEs that utilize the Inter-FE LFB 480 is expected to be IP based, which is generally assumed to be 481 congestion controlled [draft-ietf-tsvwg-rfc5405bis]. For example if 482 congestion causes a TCP packet annotated with additional ForCES 483 metadata to be dropped between FEs, the sending TCP can be expected 484 to react in the same fashion as if that packet had been dropped at a 485 different point on its path where ForCES is not involved. For this 486 reason, additional Inter-FE congestion control mechanisms are not 487 specified. 489 However, the increased packet size due to addition of ForCES metadata 490 is likely to require additional bandwidth on inter-FE links by 491 comparison to what would be required to carry the same traffic 492 without ForCES metadata. Therefore, traffic engineering SHOULD be 493 done when deploying Inter-FE encapsulation. 495 Furthermore, the Inter-FE LFB MUST only be deployed within a single 496 network (with a single network operator) or networks of an adjacent 497 set of cooperating network operators where traffic is managed to 498 avoid congestion. These are Controlled Environments, as defined by 499 Section 3.6 of [draft-ietf-tsvwg-rfc5405bis]. Additional measures 500 SHOULD be imposed to restrict the impact of Inter-FE encapsulated 501 traffic on other traffic; example: 503 o rate limiting at an upstream LFB all Inter-FE LFB traffic 505 o managed circuit breaking[circuit-b]. 507 o Isolating the Inter-FE traffic either via dedicated interfaces or 508 VLANs. 510 5.2. Inter-FE Ethernet Encapsulation 512 The Ethernet wire encapsulation is illustrated in Figure 6. The 513 process that leads to this encapsulation is described in Section 6. 514 The resulting frame is 32 bit aligned. 516 0 1 2 3 517 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 518 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 519 | Destination MAC Address | 520 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 521 | Destination MAC Address | Source MAC Address | 522 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 523 | Source MAC Address | 524 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 525 | Inter-FE ethertype | Metadata length | 526 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 527 | TLV encoded Metadata ~~~..............~~ | 528 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 529 | TLV encoded Metadata ~~~..............~~ | 530 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 531 | Original packet data ~~................~~ | 532 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 534 Figure 6: Packet format definition 536 The Ethernet header (illustrated in Figure 6) has the following 537 semantics: 539 o The Destination MAC Address is used to identify the Destination 540 FEID by the CE policy (as described in Section 6). 542 o The Source MAC Address is used to identify the Source FEID by the 543 CE policy (as described in Section 6). 545 o The Ethernet type is used to identify the frame as inter-FE LFB 546 type. Ethertype TBA1 is to be used (XXX: Note to RFC editor - 547 update when available). 549 o The 16-bit metadata length is used to described the total encoded 550 metadata length (including the 16 bits used to encode the metadata 551 length). 553 o One or more 16-bit TLV encoded Metadatum follows the metadata 554 length field. The TLV type identifies the Metadata id. ForCES 555 IANA-defined Metadata ids will be used. All TLVs will be 32 bit 556 aligned. We recognize that using a 16 bit TLV restricts the 557 metadata id to 16 bits instead of ForCES-defined component ID 558 space of 32 bits if an ILV is used. However, at the time of 559 publication we believe this is sufficient to carry all the info we 560 need; the TLV approach has been selected because it saves us 4 561 bytes per Metadatum transferred as compared to the ILV approach. 563 o The original packet data payload is appended at the end of the 564 metadata as shown. 566 6. Detailed Description of the Ethernet inter-FE LFB 568 The Ethernet inter-FE LFB has two LFB input port groups and three LFB 569 output ports as shown in Figure 7. 571 The inter-FE LFB defines two components used in aiding processing 572 described in Section 6.2. 574 +-----------------+ 575 Inter-FE LFB | | 576 Encapsulated | OUT2+--> decapsulated Packet 577 -------------->|IngressInGroup | + metadata 578 Ethernet Frame | | 579 | | 580 raw Packet + | OUT1+--> Encapsulated Ethernet 581 -------------->|EgressInGroup | Frame 582 Metadata | | 583 | EXCEPTIONOUT +--> ExceptionID, packet 584 | | + metadata 585 +-----------------+ 586 Figure 7: Inter-FE LFB 588 6.1. Data Handling 590 The Inter-FE LFB (instance) can be positioned at the egress of a 591 source FE. Figure 5 illustrates an example source FE in the form of 592 FE1. In such a case an Inter-FE LFB instance receives, via port 593 group EgressInGroup, a raw packet and associated metadata from the 594 preceding LFB instances. The input information is used to produce a 595 selection of how to generate and encapsulate the new frame. The set 596 of all selections is stored in the LFB component IFETable described 597 further below. The processed encapsulated Ethernet Frame will go out 598 on OUT1 to a downstream LFB instance when processing succeeds or to 599 the EXCEPTIONOUT port in the case of a failure. 601 The Inter-FE LFB (instance) can be positioned at the ingress of a 602 receiving FE. Figure 5 illustrates an example destination FE in the 603 form of FE1. In such a case an Inter-FE LFB receives, via an LFB 604 port in the IngressInGroup, an encapsulated Ethernet frame. 605 Successful processing of the packet will result in a raw packet with 606 associated metadata IDs going downstream to an LFB connected on OUT2. 607 On failure the data is sent out EXCEPTIONOUT. 609 6.1.1. Egress Processing 611 The egress Inter-FE LFB receives packet data and any accompanying 612 Metadatum at an LFB port of the LFB instance's input port group 613 labelled EgressInGroup. 615 The LFB implementation may use the incoming LFB port (within LFB port 616 group EgressInGroup) to map to a table index used to lookup the 617 IFETable table. 619 If lookup is successful, a matched table row which has the 620 InterFEinfo details is retrieved with the tuple {optional IFEtype, 621 optional StatId, Destination MAC address(DSTFE), Source MAC 622 address(SRCFE), optional metafilters}. The metafilters lists define 623 a whitelist of which Metadatum are to be passed to the neighboring 624 FE. The inter-FE LFB will perform the following actions using the 625 resulting tuple: 627 o Increment statistics for packet and byte count observed at 628 corresponding IFEStats entry. 630 o When MetaFilterList is present, then walk each received Metadatum 631 and apply against the MetaFilterList. If no legitimate metadata 632 is found that needs to be passed downstream then the processing 633 stops and send the packet and metadata out the EXCEPTIONOUT port 634 with exceptionID of EncapTableLookupFailed [RFC6956]. 636 o Check that the additional overhead of the Ethernet header and 637 encapsulated metadata will not exceed MTU. If it does, increment 638 the error packet count statistics and send the packet and metadata 639 out the EXCEPTIONOUT port with exceptionID of FragRequired 640 [RFC6956]. 642 o Create the Ethernet header 644 o Set the Destination MAC address of the Ethernet header with value 645 found in the DSTFE field. 647 o Set the Source MAC address of the Ethernet header with value found 648 in the SRCFE field. 650 o If the optional IFETYPE is present, set the Ethernet type to the 651 value found in IFETYPE. If IFETYPE is absent then the standard 652 Inter-FE LFB Ethernet type TBA1 is used (XXX: Note to RFC editor - 653 update when available). 655 o Encapsulate each allowed Metadatum in a TLV. Use the Metaid as 656 the "type" field in the TLV header. The TLV should be aligned to 657 32 bits. This means you may need to add padding of zeroes at the 658 end of the TLV to ensure alignment. 660 o Update the Metadata length to the sum of each TLV's space plus 2 661 bytes (for the Metadata length field 16 bit space). 663 The resulting packet is sent to the next LFB instance connected to 664 the OUT1 LFB-port; typically a port LFB. 666 In the case of a failed lookup the original packet and associated 667 metadata is sent out the EXCEPTIONOUT port with exceptionID of 668 EncapTableLookupFailed [RFC6956]. Note that the EXCEPTIONOUT LFB 669 port is merely an abstraction and implementation may in fact drop 670 packets as described above. 672 6.1.2. Ingress Processing 674 An ingressing inter-FE LFB packet is recognized by inspecting the 675 ethertype, and optionally the destination and source MAC addresses. 676 A matching packet is mapped to an LFB instance port in the 677 IngressInGroup. The IFETable table row entry matching the LFB 678 instance port may have optionally programmed metadata filters. In 679 such a case the ingress processing should use the metadata filters as 680 a whitelist of what metadatum is to be allowed. 682 o Increment statistics for packet and byte count observed. 684 o Look at the metadata length field and walk the packet data 685 extracting from the TLVs the metadata values. For each Metadatum 686 extracted, in the presence of metadata filters, the metaid is 687 compared against the relevant IFETable row metafilter list. If 688 the Metadatum is recognized, and is allowed by the filter, the 689 corresponding implementation Metadatum field is set. If an 690 unknown Metadatum id is encountered, or if the metaid is not in 691 the allowed filter list the implementation is expected to ignore 692 it, increment the packet error statistic and proceed processing 693 other Metadatum. 695 o Upon completion of processing all the metadata, the inter-FE LFB 696 instance resets the data point to the original payload (i.e skips 697 the IFE header information). At this point the original packet 698 that was passed to the egress Inter-FE LFB at the source FE is 699 reconstructed. This data is then passed along with the 700 reconstructed metadata downstream to the next LFB instance in the 701 graph. 703 In the case of processing failure of either ingress or egress 704 positioning of the LFB, the packet and metadata are sent out the 705 EXCEPTIONOUT LFB port with appropriate error id. Note that the 706 EXCEPTIONOUT LFB port is merely an abstraction and implementation may 707 in fact drop packets as described above. 709 6.2. Components 711 There are two LFB components accessed by the CE. The reader is asked 712 to refer to the definitions in Figure 8. 714 The first component, populated by the CE, is an array known as the 715 IFETable table. The array rows are made up of IFEInfo structure. 716 The IFEInfo structure constitutes: optional IFETYPE, optionally 717 present StatId, Destination MAC address(DSTFE), Source MAC 718 address(SRCFE), optionally present array of allowed Metaids 719 (MetaFilterList). 721 The second component(ID 2), populated by the FE and read by the CE, 722 is an indexed array known as the IFEStats table. Each IFEStats row 723 which carries statistics information in the structure bstats. 725 A note about the StatId relationship between the IFETable table and 726 IFEStats table: An implementation may choose to map between an 727 IFETable row and IFEStats table row using the StatId entry in the 728 matching IFETable row. In that case the IFETable StatId must be 729 present. Alternative implementation may map at provisioning time an 730 IFETable row to IFEStats table row. Yet another alternative 731 implementation may choose not to use the IFETable row StatId and 732 instead use the IFETable row index as the IFEStats index. For these 733 reasons the StatId component is optional. 735 6.3. Inter-FE LFB XML Model 737 740 742 743 PacketAny 744 Arbitrary Packet 745 746 747 InterFEFrame 748 749 Ethernet Frame with encapsulate IFE information 750 751 753 755 757 758 bstats 759 Basic stats 760 761 762 bytes 763 The total number of bytes seen 764 uint64 765 767 768 packets 769 The total number of packets seen 770 uint32 771 772 773 errors 774 The total number of packets with errors 775 uint32 776 777 779 781 782 IFEInfo 783 Describing IFE table row Information 784 785 786 IFETYPE 787 788 the ethernet type to be used for outgoing IFE frame 789 790 791 uint16 792 793 794 StatId 795 796 the Index into the stats table 797 798 799 uint32 800 801 802 DSTFE 803 804 the destination MAC address of destination FE 805 806 byte[6] 807 808 809 SRCFE 810 811 the source MAC address used for the source FE 812 813 byte[6] 814 815 816 MetaFilterList 817 818 the allowed metadata filter table 819 820 821 822 uint32 823 824 826 827 829 831 832 833 IFE 834 835 This LFB describes IFE connectivity parameterization 836 837 1.0 839 841 842 EgressInGroup 843 844 The input port group of the egress side. 845 It expects any type of Ethernet frame. 846 847 848 849 PacketAny 850 851 852 854 855 IngressInGroup 856 857 The input port group of the ingress side. 858 It expects an interFE encapsulated Ethernet frame. 859 860 861 862 InterFEFrame 863 864 865 866 868 870 871 OUT1 872 873 The output port of the egress side. 874 875 876 877 InterFEFrame 878 879 880 882 883 OUT2 884 885 The output port of the Ingress side. 886 887 888 889 PacketAny 890 891 892 894 895 EXCEPTIONOUT 896 897 The exception handling path 898 899 900 901 PacketAny 902 903 904 ExceptionID 905 906 907 909 911 913 914 IFETable 915 916 the table of all InterFE relations 917 918 919 IFEInfo 920 921 923 924 IFEStats 925 926 the stats corresponding to the IFETable table 927 928 bstats 929 931 933 934 936 938 Figure 8: Inter-FE LFB XML 940 7. Acknowledgements 942 The authors would like to thank Joel Halpern and Dave Hood for the 943 stimulating discussions. Evangelos Haleplidis shepherded and 944 contributed to improving this document. Alia Atlas was the AD 945 sponsor of this document and did a tremendous job of critiquing it. 946 The authors are grateful to Joel Halpern and Sue Hares in their roles 947 as the Routing Area reviewers in shaping the content of this 948 document. David Black put a lot of effort in making sure congestion 949 control considerations are sane. Russ Housley did the Gen-ART review 950 and Joe Touch did the TSV area. Shucheng LIU (Will) did the OPS 951 review. Suresh Krishnan helped us provide clarity during the IESG 952 review. The authors are appreciative of the efforts Stephen Farrell 953 put in fixing the security section. 955 8. IANA Considerations 957 This memo includes one IANA request within the registry https:// 958 www.iana.org/assignments/forces 959 The request is for the sub-registry "Logical Functional Block (LFB) 960 Class Names and Class Identifiers" to request for the reservation of 961 LFB class name IFE with LFB classid 18 with version 1.0. 963 +--------------+---------+---------+-------------------+------------+ 964 | LFB Class | LFB | LFB | Description | Reference | 965 | Identifier | Class | Version | | | 966 | | Name | | | | 967 +--------------+---------+---------+-------------------+------------+ 968 | 18 | IFE | 1.0 | An IFE LFB to | This | 969 | | | | standardize | document | 970 | | | | inter-FE LFB for | | 971 | | | | ForCES Network | | 972 | | | | Elements | | 973 +--------------+---------+---------+-------------------+------------+ 975 Logical Functional Block (LFB) Class Names and Class Identifiers 977 9. IEEE Assignment Considerations 979 This memo includes a request for a new ethernet protocol type as 980 described in Section 5.2. 982 10. Security Considerations 984 The FEs involved in the Inter-FE LFB belong to the same Network 985 Device (NE) and are within the scope of a single administrative 986 Ethernet LAN private network. While trust of policy in the control 987 and its treatment in the datapath exists already, an Inter-FE LFB 988 implementation SHOULD support security services provided by Media 989 Access Control Security(MACsec)[ieee8021ae]. MACsec is not currently 990 sufficiently widely deployed in traditional packet processing 991 hardware although present in newer versions of the Linux kernel 992 (which will be widely deployed) [linux-macsec]. Over time we would 993 expect that most FEs will be able to support MACsec. 995 MACsec provides security services such as message authentication 996 service and optional confidentiality service. The services can be 997 configured manually or automatically using MACsec Key Agreement(MKA) 998 over IEEE 802.1x [ieee8021x] Extensible Authentication Protocol (EAP) 999 framework. It is expected FE implementations are going to start with 1000 shared keys configured from the control plane but progress to 1001 automated key management. 1003 The following are the MACsec security mechanisms that need to be in 1004 place for the InterFE LFB: 1006 o Security mechanisms are NE-wide for all FEs. Once the security is 1007 turned on depending upon the chosen security level 1008 (Authentication, Confidentiality), it will be in effect for the 1009 inter-FE LFB for the entire duration of the session. 1011 o An operator SHOULD configure the same security policies for all 1012 participating FEs in the NE cluster. This will ensure uniform 1013 operations and avoid unnecessary complexity in policy 1014 configuration. In other words, the Security Association 1015 Keys(SAKs) should be pre-shared. When using MKA, FEs must 1016 identify themselves with a shared Connectivity Association Key 1017 (CAK) and Connectivity Association Key Name (CKN). EAP-TLS SHOULD 1018 be used as the EAP method. 1020 o An operator SHOULD configure the strict validation mode i.e all 1021 non-protected, invalid or non-verifiable frames MUST be dropped. 1023 It should be noted that given the above choices, if an FE is 1024 compromised, an entity running on the FE would be able to fake inter- 1025 FE or modify its content causing bad outcomes. 1027 11. References 1029 11.1. Normative References 1031 [RFC5810] Doria, A., Ed., Hadi Salim, J., Ed., Haas, R., Ed., 1032 Khosravi, H., Ed., Wang, W., Ed., Dong, L., Gopal, R., and 1033 J. Halpern, "Forwarding and Control Element Separation 1034 (ForCES) Protocol Specification", RFC 5810, DOI 10.17487/ 1035 RFC5810, March 2010, 1036 . 1038 [RFC5811] Hadi Salim, J. and K. Ogawa, "SCTP-Based Transport Mapping 1039 Layer (TML) for the Forwarding and Control Element 1040 Separation (ForCES) Protocol", RFC 5811, DOI 10.17487/ 1041 RFC5811, March 2010, 1042 . 1044 [RFC5812] Halpern, J. and J. Hadi Salim, "Forwarding and Control 1045 Element Separation (ForCES) Forwarding Element Model", RFC 1046 5812, DOI 10.17487/RFC5812, March 2010, 1047 . 1049 [RFC7391] Hadi Salim, J., "Forwarding and Control Element Separation 1050 (ForCES) Protocol Extensions", RFC 7391, DOI 10.17487/ 1051 RFC7391, October 2014, 1052 . 1054 [RFC7408] Haleplidis, E., "Forwarding and Control Element Separation 1055 (ForCES) Model Extension", RFC 7408, DOI 10.17487/RFC7408, 1056 November 2014, . 1058 [draft-ietf-tsvwg-rfc5405bis] 1059 Eggert, L., Fairhurst, G., and G. Shepherd, "UDP Usage 1060 Guidelines", Nov 2015, . 1063 [ieee8021ae] 1064 , "IEEE Standard for Local and metropolitan area networks 1065 Media Access Control (MAC) Security", IEEE 802.1AE-2006, 1066 Aug 2006. 1068 [ieee8021x] 1069 , "IEEE standard for local and metropolitan area networks 1070 - port-based network access control.", IEEE 802.1X-2010, 1071 2010. 1073 11.2. Informative References 1075 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 1076 Requirement Levels", BCP 14, RFC 2119, DOI 10.17487/ 1077 RFC2119, March 1997, 1078 . 1080 [RFC2460] Deering, S. and R. Hinden, "Internet Protocol, Version 6 1081 (IPv6) Specification", RFC 2460, DOI 10.17487/RFC2460, 1082 December 1998, . 1084 [RFC3746] Yang, L., Dantu, R., Anderson, T., and R. Gopal, 1085 "Forwarding and Control Element Separation (ForCES) 1086 Framework", RFC 3746, DOI 10.17487/RFC3746, April 2004, 1087 . 1089 [RFC6956] Wang, W., Haleplidis, E., Ogawa, K., Li, C., and J. 1090 Halpern, "Forwarding and Control Element Separation 1091 (ForCES) Logical Function Block (LFB) Library", RFC 6956, 1092 DOI 10.17487/RFC6956, June 2013, 1093 . 1095 [brcm-higig] 1096 , "HiGig", 1097 . 1099 [circuit-b] 1100 Fairhurst, G., "Network Transport Circuit Breakers", Feb 1101 2016, . 1104 [linux-macsec] 1105 Dubroca, S., "MACsec: Encryption for the wired LAN", 1106 netdev 11, Feb 2016. 1108 [linux-tc] 1109 Hadi Salim, J., "Linux Traffic Control Classifier-Action 1110 Subsystem Architecture", netdev 01, Feb 2015. 1112 [tc-ife] Hadi Salim, J. and D. Joachimpillai, "Distributing Linux 1113 Traffic Control Classifier-Action Subsystem", netdev 01, 1114 Feb 2015. 1116 [vxlan-udp] 1117 , "iproute2 and kernel code (drivers/net/vxlan.c)", 1118 . 1120 Authors' Addresses 1122 Damascane M. Joachimpillai 1123 Verizon 1124 60 Sylvan Rd 1125 Waltham, Mass. 02451 1126 USA 1128 Email: damascene.joachimpillai@verizon.com 1130 Jamal Hadi Salim 1131 Mojatatu Networks 1132 Suite 200, 15 Fitzgerald Rd. 1133 Ottawa, Ontario K2H 9G1 1134 Canada 1136 Email: hadi@mojatatu.com