idnits 2.17.00 (12 Aug 2021) /tmp/idnits25597/draft-ietf-bier-te-arch-07.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 (March 9, 2020) is 803 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) -- Looks like a reference, but probably isn't: '2' on line 1291 -- Looks like a reference, but probably isn't: '1' on line 1305 == Missing Reference: 'SI' is mentioned on line 1345, but not defined == Missing Reference: 'I' is mentioned on line 1352, but not defined == Missing Reference: 'VRF' is mentioned on line 1846, but not defined == Outdated reference: A later version (-06) exists of draft-ietf-bier-multicast-http-response-03 Summary: 0 errors (**), 0 flaws (~~), 5 warnings (==), 3 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group T. Eckert, Ed. 3 Internet-Draft Futurewei 4 Intended status: Standards Track G. Cauchie 5 Expires: September 10, 2020 Bouygues Telecom 6 M. Menth 7 University of Tuebingen 8 March 9, 2020 10 Tree Engineering for Bit Index Explicit Replication (BIER-TE) 11 draft-ietf-bier-te-arch-07 13 Abstract 15 This memo introduces per-packet stateless strict and loose path 16 steered replication and forwarding for Bit Index Explicit Replication 17 packets (RFC8279). This is called BIER Tree Engineering (BIER-TE). 18 BIER-TE can be used as a path steering mechanism in future Traffic 19 Engineering solutions for BIER (BIER-TE). 21 BIER-TE leverages RFC8279 and extends it with a new semantic for bits 22 in the bitstring. BIER-TE can leverage BIER forwarding engines with 23 little or no changes. 25 In BIER, the BitPositions (BP) of the packets bitstring indicate BIER 26 Forwarding Egress Routers (BFER), and hop-by-hop forwarding uses a 27 Routing Underlay such as an IGP. 29 In BIER-TE, BitPositions indicate adjacencies. The BIFT of each BFR 30 are only populated with BPs that are adjacent to the BFR in the BIER- 31 TE topology. The BIER-TE topology can consist of layer 2 or remote 32 (routed) adjacencies. The BFR then replicates and forwards BIER 33 packets to those adjacencies. This results in the aforementioned 34 strict and loose path steering and replications. 36 BIER-TE can co-exist with BIER forwarding in the same domain, for 37 example by using separate BIER sub-domains. In the absence of routed 38 adjacencies, BIER-TE does not require a BIER routing underlay, and 39 can then be operated without requiring an Interior Gateway Routing 40 protocol (IGP). 42 BIER-TE operates without explicit in-network tree-state and carries 43 the multicast distribution tree in the packet header. It can 44 therefore be a good fit to support multicast path steering in Segment 45 Routing (SR) networks. 47 Name explanation 49 [RFC-editor: This section to be removed before publication.] 51 Explanation for name change from BIER-TE to mean "Traffic 52 Engineering" to BIER-TE "Tree Engineering" in WG last-call (to 53 benefit IETF/IESG reviewers): 55 This document started by calling itself BIER-TE, "Traffic 56 Engineering" as it is a mode of BIER specifically beneficial for 57 Traffic Engineering. It supports per-packet bitstring based policy 58 steering and replication. BIER-TE technology itself does not provide 59 a complete traffic engineering solution for BIER but would require 60 combination with other technologies for a full BIER based TE 61 solution, such as a PCE and queuing mechanisms to provide bandwidth 62 and latency reservations. It is also not the only option to build a 63 traffic engineering solution utilizing BIER, for example BIER trees 64 could be steered through IGP metric engineering, such as through 65 Flex-Topologies. The architecure for Traffic Engineering with either 66 modes of BIER (BIER-TE/BIER) is intended to be defined in a separate 67 document, most likely in TEAs WG. 69 Because the name of such an overall solution is intended to be BIER- 70 TE, the expansion of BIER-TE was therefore changed to name this BIER 71 mode "Tree Engineering", so the overall solution can be distinguished 72 better from its tree building/engineering method without having to 73 change the long time well-established abbreviation BIER-TE. 75 Status of This Memo 77 This Internet-Draft is submitted in full conformance with the 78 provisions of BCP 78 and BCP 79. 80 Internet-Drafts are working documents of the Internet Engineering 81 Task Force (IETF). Note that other groups may also distribute 82 working documents as Internet-Drafts. The list of current Internet- 83 Drafts is at https://datatracker.ietf.org/drafts/current/. 85 Internet-Drafts are draft documents valid for a maximum of six months 86 and may be updated, replaced, or obsoleted by other documents at any 87 time. It is inappropriate to use Internet-Drafts as reference 88 material or to cite them other than as "work in progress." 90 This Internet-Draft will expire on September 10, 2020. 92 Copyright Notice 94 Copyright (c) 2020 IETF Trust and the persons identified as the 95 document authors. All rights reserved. 97 This document is subject to BCP 78 and the IETF Trust's Legal 98 Provisions Relating to IETF Documents 99 (https://trustee.ietf.org/license-info) in effect on the date of 100 publication of this document. Please review these documents 101 carefully, as they describe your rights and restrictions with respect 102 to this document. Code Components extracted from this document must 103 include Simplified BSD License text as described in Section 4.e of 104 the Trust Legal Provisions and are provided without warranty as 105 described in the Simplified BSD License. 107 Table of Contents 109 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 4 110 1.1. Basic Examples . . . . . . . . . . . . . . . . . . . . . 5 111 1.2. BIER-TE Topology and adjacencies . . . . . . . . . . . . 8 112 1.3. Comparison with BIER . . . . . . . . . . . . . . . . . . 9 113 1.4. Requirements Language . . . . . . . . . . . . . . . . . . 9 114 2. Components . . . . . . . . . . . . . . . . . . . . . . . . . 9 115 2.1. The Multicast Flow Overlay . . . . . . . . . . . . . . . 10 116 2.2. The BIER-TE Controller . . . . . . . . . . . . . . . . . 10 117 2.2.1. Assignment of BitPositions to adjacencies of the 118 network topology . . . . . . . . . . . . . . . . . . 11 119 2.2.2. Changes in the network topology . . . . . . . . . . . 11 120 2.2.3. Set up per-multicast flow BIER-TE state . . . . . . . 11 121 2.2.4. Link/Node Failures and Recovery . . . . . . . . . . . 12 122 2.3. The BIER-TE Forwarding Layer . . . . . . . . . . . . . . 12 123 2.4. The Routing Underlay . . . . . . . . . . . . . . . . . . 12 124 3. BIER-TE Forwarding . . . . . . . . . . . . . . . . . . . . . 12 125 3.1. The Bit Index Forwarding Table (BIFT) . . . . . . . . . . 12 126 3.2. Adjacency Types . . . . . . . . . . . . . . . . . . . . . 14 127 3.2.1. Forward Connected . . . . . . . . . . . . . . . . . . 14 128 3.2.2. Forward Routed . . . . . . . . . . . . . . . . . . . 14 129 3.2.3. ECMP . . . . . . . . . . . . . . . . . . . . . . . . 14 130 3.2.4. Local Decap . . . . . . . . . . . . . . . . . . . . . 15 131 3.3. Encapsulation considerations . . . . . . . . . . . . . . 15 132 3.4. Basic BIER-TE Forwarding Example . . . . . . . . . . . . 15 133 3.5. Forwarding comparison with BIER . . . . . . . . . . . . . 18 134 3.6. Requirements . . . . . . . . . . . . . . . . . . . . . . 18 135 4. BIER-TE Controller BitPosition Assignments . . . . . . . . . 19 136 4.1. P2P Links . . . . . . . . . . . . . . . . . . . . . . . . 19 137 4.2. BFER . . . . . . . . . . . . . . . . . . . . . . . . . . 19 138 4.3. Leaf BFERs . . . . . . . . . . . . . . . . . . . . . . . 19 139 4.4. LANs . . . . . . . . . . . . . . . . . . . . . . . . . . 20 140 4.5. Hub and Spoke . . . . . . . . . . . . . . . . . . . . . . 21 141 4.6. Rings . . . . . . . . . . . . . . . . . . . . . . . . . . 21 142 4.7. Equal Cost MultiPath (ECMP) . . . . . . . . . . . . . . . 22 143 4.8. Routed adjacencies . . . . . . . . . . . . . . . . . . . 25 144 4.8.1. Reducing BitPositions . . . . . . . . . . . . . . . . 25 145 4.8.2. Supporting nodes without BIER-TE . . . . . . . . . . 25 146 4.9. Reuse of BitPositions (without DNR) . . . . . . . . . . . 25 147 4.10. Summary of BP optimizations . . . . . . . . . . . . . . . 27 148 5. Avoiding loops and duplicates . . . . . . . . . . . . . . . . 28 149 5.1. Loops . . . . . . . . . . . . . . . . . . . . . . . . . . 28 150 5.2. Duplicates . . . . . . . . . . . . . . . . . . . . . . . 28 151 6. BIER-TE Forwarding Pseudocode . . . . . . . . . . . . . . . . 28 152 7. Managing SI, subdomains and BFR-ids . . . . . . . . . . . . . 31 153 7.1. Why SI and sub-domains . . . . . . . . . . . . . . . . . 32 154 7.2. Bit assignment comparison BIER and BIER-TE . . . . . . . 33 155 7.3. Using BFR-id with BIER-TE . . . . . . . . . . . . . . . . 33 156 7.4. Assigning BFR-ids for BIER-TE . . . . . . . . . . . . . . 34 157 7.5. Example bit allocations . . . . . . . . . . . . . . . . . 35 158 7.5.1. With BIER . . . . . . . . . . . . . . . . . . . . . . 35 159 7.5.2. With BIER-TE . . . . . . . . . . . . . . . . . . . . 36 160 7.6. Summary . . . . . . . . . . . . . . . . . . . . . . . . . 37 161 8. BIER-TE and Segment Routing . . . . . . . . . . . . . . . . . 37 162 9. Security Considerations . . . . . . . . . . . . . . . . . . . 38 163 10. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 39 164 11. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 39 165 12. Change log [RFC Editor: Please remove] . . . . . . . . . . . 39 166 13. References . . . . . . . . . . . . . . . . . . . . . . . . . 44 167 13.1. Normative References . . . . . . . . . . . . . . . . . . 44 168 13.2. Informative References . . . . . . . . . . . . . . . . . 44 169 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 45 171 1. Introduction 173 BIER-TE shares architecture, terminology and packet formats with BIER 174 as described in [RFC8279] and [RFC8296]. This document describes 175 BIER-TE in the expectation that the reader is familiar with these two 176 documents. 178 In BIER-TE, BitPositions (BP) indicate adjacencies. The BIFT of each 179 BFR is only populated with BP that are adjacent to the BFR in the 180 BIER-TE Topology. Other BPs are left without adjacency. The BFR 181 replicate and forwards BIER packets to adjacent BPs that are set in 182 the packet. BPs are normally also reset upon forwarding to avoid 183 duplicates and loops. This is detailed further below. 185 Note that related work, [I-D.ietf-roll-ccast] uses bloom filters to 186 represent leaves or edges of the intended delivery tree. Bloom 187 filters in general can support larger trees/topologies with fewer 188 addressing bits than explicit bitstrings, but they introduce the 189 heuristic risk of false positives and cannot reset bits in the 190 bitstring during forwarding to avoid loops. For these reasons, BIER- 191 TE uses explicit bitstrings like BIER. The explicit bitstrings of 192 BIER-TE can also be seen as a special type of bloom filter, and this 193 is how related work [ICC] describes it. 195 1.1. Basic Examples 197 BIER-TE forwarding is best introduced with simple examples. 199 BIER-TE Topology: 201 Diagram: 203 p5 p6 204 --- BFR3 --- 205 p3/ p13 \p7 206 BFR1 ---- BFR2 BFR5 ----- BFR6 207 p1 p2 p4\ p14 /p10 p11 p12 208 --- BFR4 --- 209 p8 p9 211 (simplified) BIER-TE Bit Index Forwarding Tables (BIFT): 213 BFR1: p1 -> local_decap 214 p2 -> forward_connected to BFR2 216 BFR2: p1 -> forward_connected to BFR1 217 p5 -> forward_connected to BFR3 218 p8 -> forward_connected to BFR4 220 BFR3: p3 -> forward_connected to BFR2 221 p7 -> forward_connected to BFR5 222 p13 -> local_decap 224 BFR4: p4 -> forward_connected to BFR2 225 p10 -> forward_connected to BFR5 226 p14 -> local_decap 228 BFR5: p6 -> forward_connected to BFR3 229 p9 -> forward_connected to BFR4 230 p12 -> forward_connected to BFR6 232 BFR6: p11 -> forward_connected to BFR5 233 p12 -> local_decap 235 Figure 1: BIER-TE basic example 237 Consider the simple network in the above BIER-TE overview example 238 picture with 6 BFRs. p1...p14 are the BitPositions (BP) used. All 239 BFRs can act as ingress BFR (BFIR), BFR1, BFR3, BFR4 and BFR6 can 240 also be egress BFR (BFER). Forward_connected is the name for 241 adjacencies that are representing subnet adjacencies of the network. 242 Local_decap is the name of the adjacency to decapsulate BIER-TE 243 packets and pass their payload to higher layer processing. 245 Assume a packet from BFR1 should be sent via BFR4 to BFR6. This 246 requires a bitstring (p2,p8,p10,p12). When this packet is examined 247 by BIER-TE on BFR1, the only BitPosition from the bitstring that is 248 also set in the BIFT is p2. This will cause BFR1 to send the only 249 copy of the packet to BFR2. Similarly, BFR2 will forward to BFR4 250 because of p8, BFR4 to BFR5 because of p10 and BFR5 to BFR6 because 251 of p12. p12 also makes BFR6 receive and decapsulate the packet. 253 To send in addition to BFR6 via BFR4 also a copy to BFR3, the 254 bitstring needs to be (p2,p5,p8,p10,p12,p13). When this packet is 255 examined by BFR2, p5 causes one copy to be sent to BFR3 and p8 one 256 copy to BFR4. When BFR3 receives the packet, p13 will cause it to 257 receive and decapsulate the packet. 259 If instead the bitstring was (p2,p6,p8,p10,p12,p13), the packet would 260 be copied by BFR5 towards BFR3 because p6 instead of BFR2 to BFR5 261 because of p6 in the prior case. This is showing the ability of the 262 shown BIER-TE Topology to make the traffic pass across any possible 263 path and be replicated where desired. 265 BIER-TE has various options to minimize BP assignments, many of which 266 are based on assumptions about the required multicast traffic paths 267 and bandwidth consumption in the network. 269 The following picture shows a modified example, in which Rtr2 and 270 Rtr5 are assumed not to support BIER-TE, so traffic has to be unicast 271 encapsulated across them. Unicast tunneling of BIER-TE packets can 272 leverage any feasible mechanism such as MPLS or IP, these 273 encapsulations are out of scope of this document. To emphasize non- 274 native forwarding of BIER-TE packets, these adjacencies are called 275 "forward_routed", but otherwise there is no difference in their 276 processing over the aforementioned "forward_connected" adjacencies. 278 In addition, bits are saved in the following example by assuming that 279 BFR1 only needs to be BFIR but not BFER or transit BFR. 281 BIER-TE Topology: 283 Diagram: 285 p1 p3 p7 286 ....> BFR3 <.... p5 287 ........ ........> 288 BFR1 (Rtr2) (Rtr5) BFR6 289 ........ ........> 290 ....> BFR4 <.... p6 291 p2 p4 p8 293 (simplified) BIER-TE Bit Index Forwarding Tables (BIFT): 295 BFR1: p1 -> forward_routed to BFR3 296 p2 -> forward_routed to BFR4 298 BFR3: p3 -> local_decap 299 p5 -> forward_routed to BFR6 301 BFR4: p4 -> local_decap 302 p6 -> forward_routed to BFR6 304 BFR6: p5 -> local_decap 305 p6 -> local_decap 306 p7 -> forward_routed to BFR3 307 p8 -> forward_routed to BFR4 309 Figure 2: BIER-TE basic overlay example 311 To send a BIER-TE packet from BFR1 via BFR3 to BFR6, the bitstring is 312 (p1,p5). From BFR1 via BFR4 to BFR6 it is (p2,p6). A packet from 313 BFR1 to BFR3,BFR4 and BFR6 can use (p1,p2,p3,p4,p5) or 314 (p1,p2,p3,p4,p6), or via BFR6 (p2,p3,p4,p6,p7) or (p1.p3,p4,p5,p8). 316 1.2. BIER-TE Topology and adjacencies 318 The key new component in BIER-TE to control where replication can or 319 should happens and how to minimize the required BP for segments is - 320 as shown in these two examples - the BIER-TE topology. 322 The BIER-TE Topology consists of the BIFT of all the BFR and can also 323 be expressed in a diagram as a graph where the edges are the 324 adjacencies between the BFR. Adjacencies are naturally 325 unidirectional. BP can be reused across multiple adjacencies as long 326 as this does not lead to undesired duplicates or loops as explained 327 further down in the text. 329 If the BIER-TE topology represents the underlying (layer 2) topology 330 of the network, this is called "native" BIER-TE as shown in the first 331 example. This can be freely mixed with "overlay" BIER-TE, in 332 "forward_routed" adjacencies are used. 334 1.3. Comparison with BIER 336 The key differences over BIER are: 338 o BIER-TE replaces in-network autonomous path calculation by 339 explicit paths calculated off-path by the BIER-TE Controller. 341 o In BIER-TE every BitPosition of the BitString of a BIER-TE packet 342 indicates one or more adjacencies - instead of a BFER as in BIER. 344 o BIER-TE in each BFR has no routing table but only a BIER-TE 345 Forwarding Table (BIFT) indexed by SI:BitPosition and populated 346 with only those adjacencies to which the BFR should replicate 347 packets to. 349 BIER-TE headers use the same format as BIER headers. 351 BIER-TE forwarding does not require/use the BFIR-ID. The BFIR-ID can 352 still be useful though for coordinated BFIR/BFER functions, such as 353 the context for upstream assigned labels for MPLS payloads in MVPN 354 over BIER-TE. 356 If the BIER-TE domain is also running BIER, then the BFIR-ID in BIER- 357 TE packets can be set to the same BFIR-ID as used with BIER packets. 359 If the BIER-TE domain is not running full BIER or does not want to 360 reduce the need to allocate bits in BIER bitstrings for BFIR-ID 361 values, then the allocation of BFIR-ID values in BIER-TE packets can 362 be done through other mechanisms outside the scope of this document, 363 as long as this is appropriately agreed upon between all BFIR/BFER. 365 1.4. Requirements Language 367 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 368 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 369 document are to be interpreted as described in RFC 2119 [RFC2119]. 371 2. Components 373 End to end BIER-TE operations consists of four mayor components: The 374 "Multicast Flow Overlay", the "BIER-TE control plane" consisting of 375 the "BIER-TE Controller" and its signaling channels to the BFR, the 376 "Routing Underlay" and the "BIER-TE forwarding layer". The Bier-TE 377 Controller is the new architectural component in BIER-TE compared to 378 BIER. 380 Picture 2: Components of BIER-TE 382 <------BGP/PIM-----> 383 |<-IGMP/PIM-> multicast flow <-PIM/IGMP->| 384 overlay 386 [BIER-TE Controller] <=> [BIER-TE Topology] 387 BIER-TE control plane 388 ^ ^ ^ 389 / | \ BIER-TE control protocol 390 | | | e.g. Netconf/Restconf/Yang 391 v v v 392 Src -> Rtr1 -> BFIR-----BFR-----BFER -> Rtr2 -> Rcvr 394 |<----------------->| 395 BIER-TE forwarding layer 397 |<- BIER-TE domain->| 399 |<--------------------->| 400 Routing underlay 402 Figure 3: BIER-TE architecture 404 2.1. The Multicast Flow Overlay 406 The Multicast Flow Overlay operates as in BIER. See [RFC8279]. 407 Instead of interacting with the BIER forwarding layer (as in BIER), 408 it interacts with the BIER-TE Controller. 410 2.2. The BIER-TE Controller 412 The BIER-TE Controller is representing the control plane of BIER-TE. 413 It communicates two sets of information with BFRs: 415 During initial provisioning or modifications of the network topology, 416 the BIER-TE Controller discovers the network topology and creates the 417 BIER-TE topology from it: determine which adjacencies are required/ 418 desired and assign BitPositions to them. Then it signals the 419 resulting of BitPositions and their adjacencies to each BFR to set up 420 their BIER-TE BIFTs. 422 During day-to-day operations of the network, the BIER-TE Controller 423 signals to BFIRs what multicast flows are mapped to what BitStrings. 425 Communications between the BIER-TE Controller and BFRs is ideally via 426 standardized protocols and data-models such as Netconf/Restconf/Yang. 427 This is currently outside the scope of this document. Vendor- 428 specific CLI on the BFRs is also a possible stopgap option (as in 429 many other SDN solutions lacking definition of standardized data 430 model). 432 For simplicity, the procedures of the BIER-TE Controller are 433 described in this document as if it is a single, centralized 434 automated entity, such as an SDN controller. It could equally be an 435 operator setting up CLI on the BFRs. Distribution of the functions 436 of the BIER-TE Controller is currently outside the scope of this 437 document. 439 2.2.1. Assignment of BitPositions to adjacencies of the network 440 topology 442 The BIER-TE Controller tracks the BFR topology of the BIER-TE domain. 443 It determines what adjacencies require BitPositions so that BIER-TE 444 explicit paths can be built through them as desired by operator 445 policy. 447 The BIER-TE Controller then pushes the BitPositions/adjacencies to 448 the BIFT of the BFRs, populating only those SI:BitPositions to the 449 BIFT of each BFR to which that BFR should be able to send packets to 450 - adjacencies connecting to this BFR. 452 2.2.2. Changes in the network topology 454 If the network topology changes (not failure based) so that 455 adjacencies that are assigned to BitPositions are no longer needed, 456 the BIER-TE Controller can re-use those BitPositions for new 457 adjacencies. First, these BitPositions need to be removed from any 458 BFIR flow state and BFR BIFT state, then they can be repopulated, 459 first into BIFT and then into the BFIR. 461 2.2.3. Set up per-multicast flow BIER-TE state 463 The BIER-TE Controller interacts with the multicast flow overlay to 464 determine what multicast flow needs to be sent by a BFIR to which set 465 of BFER. It calculates the desired distribution tree across the 466 BIER-TE domain based on algorithms outside the scope of this document 467 (e.g. CSFP, Steiner Tree, ...). It then pushes the calculated 468 BitString into the BFIR. 470 See [I-D.ietf-bier-multicast-http-response] for a solution describing 471 this interaction. 473 2.2.4. Link/Node Failures and Recovery 475 When link or nodes fail or recover in the topology, BIER-TE can 476 quickly respond with the optional FRR procedures described in [I- 477 D.eckert-bier-te-frr]. It can also more slowly react by 478 recalculating the BitStrings of affected multicast flows. This 479 reaction is slower than the FRR procedure because the BIER-TE 480 Controller needs to receive link/node up/down indications, 481 recalculate the desired BitStrings and push them down into the BFIRs. 482 With FRR, this is all performed locally on a BFR receiving the 483 adjacency up/down notification. 485 2.3. The BIER-TE Forwarding Layer 487 When the BIER-TE Forwarding Layer receives a packet, it simply looks 488 up the BitPositions that are set in the BitString of the packet in 489 the Bit Index Forwarding Table (BIFT) that was populated by the BIER- 490 TE Controller. For every BP that is set in the BitString, and that 491 has one or more adjacencies in the BIFT, a copy is made according to 492 the type of adjacencies for that BP in the BIFT. Before sending any 493 copy, the BFR resets all BP in the BitString of the packet for which 494 the BFR has one or more adjacencies in the BIFT, except when the 495 adjacency indicates "DoNotReset" (DNR, see Section 3.2.1). This is 496 done to inhibit that packets can loop. 498 2.4. The Routing Underlay 500 BIER-TE is sending BIER packets to directly connected BIER-TE 501 neighbors as L2 (unicasted) BIER packets without requiring a routing 502 underlay. BIER-TE forwarding uses the Routing underlay for 503 forward_routed adjacencies which copy BIER-TE packets to not- 504 directly-connected BFRs (see below for adjacency definitions). 506 If the BFR intends to support FRR for BIER-TE, then the BIER-TE 507 forwarding plane needs to receive fast adjacency up/down 508 notifications: Link up/down or neighbor up/down, e.g. from BFD. 509 Providing these notifications is considered to be part of the routing 510 underlay in this document. 512 3. BIER-TE Forwarding 514 3.1. The Bit Index Forwarding Table (BIFT) 516 The Bit Index Forwarding Table (BIFT) exists in every BFR. For every 517 subdomain in use, it is a table indexed by SI:BitPosition and is 518 populated by the BIER-TE control plane. Each index can be empty or 519 contain a list of one or more adjacencies. 521 BIER-TE can support multiple subdomains like BIER. Each one with a 522 separate BIFT 524 In the BIER architecture, indices into the BIFT are explained to be 525 both BFR-id and SI:BitString (BitPosition). This is because there is 526 a 1:1 relationship between BFR-id and SI:BitString - every bit in 527 every SI is/can be assigned to a BFIR/BFER. In BIER-TE there are 528 more bits used in each BitString than there are BFIR/BFER assigned to 529 the bitstring. This is because of the bits required to express the 530 engineered path through the topology. The BIER-TE forwarding 531 definitions do therefore not use the term BFR-id at all. Instead, 532 BFR-ids are only used as required by routing underlay, flow overlay 533 of BIER headers. Please refer to Section 7 for explanations how to 534 deal with SI, subdomains and BFR-id in BIER-TE. 536 ------------------------------------------------------------------ 537 | Index: | Adjacencies: | 538 | SI:BitPosition | or one or more per entry | 539 ================================================================== 540 | 0:1 | forward_connected(interface,neighbor{,DNR}) | 541 ------------------------------------------------------------------ 542 | 0:2 | forward_connected(interface,neighbor{,DNR}) | 543 | | forward_connected(interface,neighbor{,DNR}) | 544 ------------------------------------------------------------------ 545 | 0:3 | local_decap({VRF}) | 546 ------------------------------------------------------------------ 547 | 0:4 | forward_routed({VRF,}l3-neighbor) | 548 ------------------------------------------------------------------ 549 | 0:5 | | 550 ------------------------------------------------------------------ 551 | 0:6 | ECMP({adjacency1,...adjacencyN}, seed) | 552 ------------------------------------------------------------------ 553 ... 554 | BitStringLength | ... | 555 ------------------------------------------------------------------ 556 Bit Index Forwarding Table 558 Figure 4: BIFT adjacencies 560 The BIFT is programmed into the data plane of BFRs by the BIER-TE 561 Controller and used to forward packets, according to the rules 562 specified in the BIER-TE Forwarding Procedures. 564 Adjacencies for the same BP when populated in more than one BFR by 565 the BIER-TE Controller does not have to have the same adjacencies. 566 This is up to the BIER-TE Controller. BPs for p2p links are one case 567 (see below). 569 3.2. Adjacency Types 571 3.2.1. Forward Connected 573 A "forward_connected" adjacency is towards a directly connected BFR 574 neighbor using an interface address of that BFR on the connecting 575 interface. A forward_connected adjacency does not route packets but 576 only L2 forwards them to the neighbor. 578 Packets sent to an adjacency with "DoNotReset" (DNR) set in the BIFT 579 will not have the BitPosition for that adjacency reset when the BFR 580 creates a copy for it. The BitPosition will still be reset for 581 copies of the packet made towards other adjacencies. This can be 582 used for example in ring topologies as explained below. 584 3.2.2. Forward Routed 586 A "forward_routed" adjacency is an adjacency towards a BFR that is 587 not a forward_connected adjacency: towards a loopback address of a 588 BFR or towards an interface address that is non-directly connected. 589 Forward_routed packets are forwarded via the Routing Underlay. 591 If the Routing Underlay has multiple paths for a forward_routed 592 adjacency, it will perform ECMP independent of BIER-TE for packets 593 forwarded across a forward_routed adjacency. This is independent of 594 BIER-TE ECMP described in Section 3.2.3. 596 If the Routing Underlay has FRR, it will perform FRR independent of 597 BIER-TE for packets forwarded across a forward_routed adjacency. 599 3.2.3. ECMP 601 The ECMP mechanisms in BIER are tied to the BIER BIFT and are 602 therefore not directly useable with BIER-TE. The following 603 procedures describe ECMP for BIER-TE that we consider to be 604 lightweight but also well manageable. It leverages the existing 605 entropy parameter in the BIER header to keep packets of the flows on 606 the same path and it introduces a "seed" parameter to allow for 607 traffic flows to be polarized or randomized across multiple hops. 609 An "Equal Cost Multipath" (ECMP) adjacency has a list of two or more 610 adjacencies included in it. It copies the BIER-TE to one of those 611 adjacencies based on the ECMP hash calculation. The BIER-TE ECMP 612 hash algorithm must select the same adjacency from that list for all 613 packets with the same "entropy" value in the BIER-TE header if the 614 same number of adjacencies and same seed are given as parameters. 615 Further use of the seed parameter is explained below. 617 3.2.4. Local Decap 619 A "local_decap" adjacency passes a copy of the payload of the BIER-TE 620 packet to the packets NextProto within the BFR (IPv4/IPv6, 621 Ethernet,...). A local_decap adjacency turns the BFR into a BFER for 622 matching packets. Local_decap adjacencies require the BFER to 623 support routing or switching for NextProto to determine how to 624 further process the packet. 626 3.3. Encapsulation considerations 628 Specifications for BIER-TE encapsulation are outside the scope of 629 this document. This section gives explanations and guidelines. 631 Because a BFR needs to interpret the BitString of a BIER-TE packet 632 differently from a BIER packet, it is necessary to distinguish BIER 633 from BIER-TE packets. This is subject to definitions in BIER 634 encapsulation specifications. 636 MPLS encapsulation [RFC8296] for example assigns one label by which 637 BFRs recognizes BIER packets for every (SI,subdomain) combination. 638 If it is desirable that every subdomain can forward only BIER or 639 BIER-TE packets, then the label allocation could stay the same, and 640 only the forwarding model (BIER/BIER-TE) would have to be defined per 641 subdomain. If it is desirable to support both BIER and BIER-TE 642 forwarding in the same subdomain, then additional labels would need 643 to be assigned for BIER-TE forwarding. 645 "forward_routed" requires an encapsulation permitting to unicast 646 BIER-TE packets to a specific interface address on a target BFR. 647 With MPLS encapsulation, this can simply be done via a label stack 648 with that addresses label as the top label - followed by the label 649 assigned to (SI,subdomain) - and if necessary (see above) BIER-TE. 650 With non-MPLS encapsulation, some form of IP tunneling (IP in IP, 651 LISP, GRE) would be required. 653 The encapsulation used for "forward_routed" adjacencies can equally 654 support existing advanced adjacency information such as "loose source 655 routes" via e.g. MPLS label stacks or appropriate header extensions 656 (e.g. for IPv6). 658 3.4. Basic BIER-TE Forwarding Example 660 [RFC Editor: remove this section.] 662 THIS SECTION TO BE REMOVED IN RFC BECAUSE IT WAS SUPERCEEDED BY 663 SECTION 1.1 EXAMPLE - UNLESS REVIEWERS CHIME IN AND EXPRESS DESIRE TO 664 KEEP THIS ADDITIONAL EXAMPLE SECTION. 666 Step by step example of basic BIER-TE forwarding. This does not use 667 ECMP or forward_routed adjacencies nor does it try to minimize the 668 number of required BitPositions for the topology. 670 [BIER-TE Controller] 671 / | \ 672 v v v 674 | p13 p1 | 675 +- BFIR2 --+ | 676 | | p2 p6 | LAN2 677 | +-- BFR3 --+ | 678 | | | p7 p11 | 679 Src -+ +-- BFER1 --+ 680 | | p3 p8 | | 681 | +-- BFR4 --+ +-- Rcv1 682 | | | | 683 | | 684 | p14 p4 | 685 +- BFIR1 --+ | 686 | +-- BFR5 --+ p10 p12 | 687 LAN1 | p5 p9 +-- BFER2 --+ 688 | +-- Rcv2 689 | 690 LAN3 692 IP |..... BIER-TE network......| IP 694 Figure 5: BIER-TE Forwarding Example 696 pXX indicate the BitPositions number assigned by the BIER-TE 697 Controller to adjacencies in the BIER-TE topology. For example, p9 698 is the adjacency towards BFR5 on the LAN connecting to BFER2. 700 BIFT BFIR2: 701 p13: local_decap() 702 p2: forward_connected(BFR3) 704 BIFT BFR3: 705 p1: forward_connected(BFIR2) 706 p7: forward_connected(BFER1) 707 p8: forward_connected(BFR4) 709 BIFT BFER1: 710 p11: local_decap() 711 p6: forward_connected(BFR3) 712 p8: forward_connected(BFR4) 714 Figure 6: BIER-TE Forwarding Example Adjacencies 716 ...and so on. 718 For example, we assume that some multicast traffic seen on LAN1 needs 719 to be sent via BIER-TE by BFIR2 towards Rcv1 and Rcv2. The BIER-TE 720 Controller determines it wants it to pass this traffic across the 721 following paths: 723 -> BFER1 ---------------> Rcv1 724 BFIR2 -> BFR3 725 -> BFR4 -> BFR5 -> BFER2 -> Rcv2 727 Figure 7: BIER-TE Forwarding Example Paths 729 These paths equal to the following BitString: p2, p5, p7, p8, p10, 730 p11, p12. 732 This BitString is assigned by BFIR2 to the example multicast traffic 733 received from LAN1. 735 Then BFIR2 forwards this multicast traffic with BIER-TE based on that 736 BitString. The BIFT of BFIR2 has only p2 and p13 populated. Only p2 737 is in the BitString and this is an adjacency towards BFR3. BFIR2 738 therefore resets p2 in the BitString and sends a copy towards BFR2. 740 BFR3 sees a BitString of p5,p7,p8,p10,p11,p12. It is only interested 741 in p1,p7,p8. It creates a copy of the packet to BFER1 (due to p7) 742 and one to BFR4 (due to p8). It resets p7, p8 before sending. 744 BFER1 sees a BitString of p5,p10,p11,p12. It is only interested in 745 p6,p7,p8,p11 and therefore considers only p11. p11 is a "local_decap" 746 adjacency installed by the BIER-TE Controller because BFER1 should 747 pass packets to IP multicast. The local_decap adjacency instructs 748 BFER1 to create a copy, decapsulate it from the BIER header and pass 749 it on to the NextProtocol, in this example IP multicast. IP 750 multicast will then forward the packet out to LAN2 because it did 751 receive PIM or IGMP joins on LAN2 for the traffic. 753 Further processing of the packet in BFR4, BFR5 and BFER2 accordingly. 755 3.5. Forwarding comparison with BIER 757 Forwarding of BIER-TE is designed to allow common forwarding hardware 758 with BIER. In fact, one of the main goals of this document is to 759 encourage the building of forwarding hardware that can not only 760 support BIER, but also BIER-TE - to allow experimentation with BIER- 761 TE and support building of BIER-TE control plane code. 763 The pseudocode in Section 6 shows how existing BIER/BIFT forwarding 764 can be amended to support basic BIER-TE forwarding, by using BIER 765 BIFT's F-BM. Only the masking of bits due to avoid duplicates must 766 be skipped when forwarding is for BIER-TE. 768 Whether to use BIER or BIER-TE forwarding can simply be a configured 769 choice per subdomain and accordingly be set up by a BIER-TE 770 Controller. The BIER packet encapsulation [RFC8296] too can be 771 reused without changes except that the currently defined BIER-TE ECMP 772 adjacency does not leverage the entropy field so that field would be 773 unused when BIER-TE forwarding is used. 775 3.6. Requirements 777 Basic BIER-TE forwarding MUST support to configure Subdomains to use 778 basic BIER-TE forwarding rules (instead of BIER). With basic BIER-TE 779 forwarding, every bit MUST support to have zero or one adjacency. It 780 MUST support the adjacency types forward_connected without DNR flag, 781 forward_routed and local_decap. All other BIER-TE forwarding 782 features are optional. These basic BIER-TE requirements make BIER-TE 783 forwarding exactly the same as BIER forwarding with the exception of 784 skipping the aforementioned F-BM masking on egress. 786 BIER-TE forwarding SHOULD support the DNR flag, as this is highly 787 useful to save bits in rings (see Section 4.6). 789 BIER-TE forwarding MAY support more than one adjacency on a bit and 790 ECMP adjacencies. The importance of ECMP adjacencies is unclear when 791 traffic steering is used because it may be more desirable to 792 explicitly steer traffic across non-ECMP paths to make per-path 793 traffic calculation easier for BIER-TE Controllers. Having more than 794 one adjacency for a bit allows further savings of bits in hub&spoke 795 scenarios, but unlike rings it is less "natural" to flood traffic 796 across multiple links unconditional. Both ECMP and multiple 797 adjacencies are forwarding plane features that should be possible to 798 support later when needed as they do not impact the basic BIER-TE 799 replication loop. This is true because there is no inter-copy 800 dependency through resetting of F-BM as in BIER. 802 4. BIER-TE Controller BitPosition Assignments 804 This section describes how the BIER-TE Controller can use the 805 different BIER-TE adjacency types to define the BitPositions of a 806 BIER-TE domain. 808 Because the size of the BitString is limiting the size of the BIER-TE 809 domain, many of the options described exist to support larger 810 topologies with fewer BitPositions (4.1, 4.3, 4.4, 4.5, 4.6, 4.7, 811 4.8). 813 4.1. P2P Links 815 Each P2p link in the BIER-TE domain is assigned one unique 816 BitPosition with a forward_connected adjacency pointing to the 817 neighbor on the p2p link. 819 4.2. BFER 821 Every non-Leaf BFER is given a unique BitPosition with a local_decap 822 adjacency. 824 4.3. Leaf BFERs 826 BFR1(P) BFR2(P) BFR1(P) BFR2(P) 827 | \ / | | | 828 | X | | | 829 | / \ | | | 830 BFER1(PE) BFER2(PE) BFER1(PE)----BFER2(PE) 832 Leaf BFER / Non-Leaf BFER / 833 PE-router PE-router 835 Figure 8: Leaf vs. non-Leaf BFER Example 837 Leaf BFERs are BFERs where incoming BIER-TE packets never need to be 838 forwarded to another BFR but are only sent to the BFER to exit the 839 BIER-TE domain. For example, in networks where PEs are spokes 840 connected to P routers, those PEs are Leaf BFERs unless there is a 841 U-turn between two PEs. Consider how redundant disjoint traffic can 842 reach BFER1/BFER2 in above picture: When BFER1/BFER2 are Non-Leaf 843 BFER as shown on the right hand side, one traffic copy would be 844 forwarded to BFER1 from BFR1, but the other one could only reach 845 BFER1 via BFER2, which makes BFER2 a non-Leaf BFER. Likewise BFER1 846 is a non-Leaf BFER when forwarding traffic to BFER2. 848 Note that the BFERs in the left hand picture are only guaranteed to 849 be leaf-BFER by fitting routing configuration that prohibits transit 850 traffic to pass through the BFERs, which is commonly applied in these 851 topologies. 853 All leaf-BFER in a BIER-TE domain can share a single BitPosition. 854 This is possible because the BitPosition for the adjacency to reach 855 the BFER can be used to distinguish whether or not packets should 856 reach the BFER. 858 This optimization will not work if an upstream interface of the BFER 859 is using a BitPosition optimized as described in the following two 860 sections (LAN, Hub and Spoke). 862 4.4. LANs 864 In a LAN, the adjacency to each neighboring BFR on the LAN is given a 865 unique BitPosition. The adjacency of this BitPosition is a 866 forward_connected adjacency towards the BFR and this BitPosition is 867 populated into the BIFT of all the other BFRs on that LAN. 869 BFR1 870 |p1 871 LAN1-+-+---+-----+ 872 p3| p4| p2| 873 BFR3 BFR4 BFR7 875 Figure 9: LAN Example 877 If Bandwidth on the LAN is not an issue and most BIER-TE traffic 878 should be copied to all neighbors on a LAN, then BitPositions can be 879 saved by assigning just a single BitPosition to the LAN and 880 populating the BitPosition of the BIFTs of each BFRs on the LAN with 881 a list of forward_connected adjacencies to all other neighbors on the 882 LAN. 884 This optimization does not work in the case of BFRs redundantly 885 connected to more than one LANs with this optimization because these 886 BFRs would receive duplicates and forward those duplicates into the 887 opposite LANs. Adjacencies of such BFRs into their LANs still need a 888 separate BitPosition. 890 4.5. Hub and Spoke 892 In a setup with a hub and multiple spokes connected via separate p2p 893 links to the hub, all p2p links can share the same BitPosition. The 894 BitPosition on the hub's BIFT is set up with a list of 895 forward_connected adjacencies, one for each Spoke. 897 This option is similar to the BitPosition optimization in LANs: 898 Redundantly connected spokes need their own BitPositions. 900 This type of optimized BP could be used for example when all traffic 901 is "broadcast" traffic (very dense receiver set) such as live-TV or 902 situation-awareness (SA). This BP optimization can then be used to 903 explicitly steer different traffic flows across different ECMP paths 904 in Data-Center or broadband-aggregation networks with minimal use of 905 BPs. 907 4.6. Rings 909 In L3 rings, instead of assigning a single BitPosition for every p2p 910 link in the ring, it is possible to save BitPositions by setting the 911 "Do Not Reset" (DNR) flag on forward_connected adjacencies. 913 For the rings shown in the following picture, a single BitPosition 914 will suffice to forward traffic entering the ring at BFRa or BFRb all 915 the way up to BFR1: 917 On BFRa, BFRb, BFR30,... BFR3, the BitPosition is populated with a 918 forward_connected adjacency pointing to the clockwise neighbor on the 919 ring and with DNR set. On BFR2, the adjacency also points to the 920 clockwise neighbor BFR1, but without DNR set. 922 Handling DNR this way ensures that copies forwarded from any BFR in 923 the ring to a BFR outside the ring will not have the ring BitPosition 924 set, therefore minimizing the chance to create loops. 926 v v 927 | | 928 L1 | L2 | L3 929 /-------- BFRa ---- BFRb --------------------\ 930 | | 931 \- BFR1 - BFR2 - BFR3 - ... - BFR29 - BFR30 -/ 932 | | L4 | | 933 p33| p15| 934 BFRd BFRc 936 Figure 10: Ring Example 938 Note that this example only permits for packets to enter the ring at 939 BFRa and BFRb, and that packets will always travel clockwise. If 940 packets should be allowed to enter the ring at any ring BFR, then one 941 would have to use two ring BitPositions. One for clockwise, one for 942 counterclockwise. 944 Both would be set up to stop rotating on the same link, e.g. L1. 945 When the ingress ring BFR creates the clockwise copy, it will reset 946 the counterclockwise BitPosition because the DNR bit only applies to 947 the bit for which the replication is done. Likewise for the 948 clockwise BitPosition for the counterclockwise copy. In result, the 949 ring ingress BFR will send a copy in both directions, serving BFRs on 950 either side of the ring up to L1. 952 4.7. Equal Cost MultiPath (ECMP) 954 The ECMP adjacency allows to use just one BP per link bundle between 955 two BFRs instead of one BP for each p2p member link of that link 956 bundle. In the following picture, one BP is used across L1,L2,L3. 958 --L1----- 959 BFR1 --L2----- BFR2 960 --L3----- 962 BIFT entry in BFR1: 963 ------------------------------------------------------------------ 964 | Index | Adjacencies | 965 ================================================================== 966 | 0:6 | ECMP({forward_connected(L1, BFR2), | 967 | | forward_connected(L2, BFR2), | 968 | | forward_connected(L3, BFR2)}, seed) | 969 ------------------------------------------------------------------ 971 BIFT entry in BFR2: 972 ------------------------------------------------------------------ 973 | Index | Adjacencies | 974 ================================================================== 975 | 0:6 | ECMP({forward_connected(L1, BFR1), | 976 | | forward_connected(L2, BFR1), | 977 | | forward_connected(L3, BFR1)}, seed) | 978 ------------------------------------------------------------------ 980 Figure 11: ECMP Example 982 This document does not standardize any ECMP algorithm because it is 983 sufficient for implementations to document their freely chosen ECMP 984 algorithm. This allows the BIER-TE Controller to calculate ECMP 985 paths and seeds. The following picture shows an example ECMP 986 algorithm: 988 forward(packet, ECMP(adj(0), adj(1),... adj(N-1), seed)): 989 i = (packet(bier-header-entropy) XOR seed) % N 990 forward packet to adj(i) 992 Figure 12: ECMP algorithm Example 994 In the following example, all traffic from BFR1 towards BFR10 is 995 intended to be ECMP load split equally across the topology. This 996 example is not meant as a likely setup, but to illustrate that ECMP 997 can be used to share BPs not only across link bundles, and it 998 explains the use of the seed parameter. 1000 BFR1 (BFIR) 1001 /L11 \L12 1002 / \ 1003 BFR2 BFR3 1004 /L21 \L22 /L31 \L32 1005 / \ / \ 1006 BFR4 BFR5 BFR6 BFR7 1007 \ / \ / 1008 \ / \ / 1009 BFR8 BFR9 1010 \ / 1011 \ / 1012 BFR10 (BFER) 1014 BIFT entry in BFR1: 1015 ------------------------------------------------------------------ 1016 | 0:6 | ECMP({forward_connected(L11, BFR2), | 1017 | | forward_connected(L12, BFR3)}, seed1) | 1018 ------------------------------------------------------------------ 1020 BIFT entry in BFR2: 1021 ------------------------------------------------------------------ 1022 | 0:7 | ECMP({forward_connected(L21, BFR4), | 1023 | | forward_connected(L22, BFR5)}, seed1) | 1024 ------------------------------------------------------------------ 1026 BIFT entry in BFR3: 1027 ------------------------------------------------------------------ 1028 | 0:7 | ECMP({forward_connected(L31, BFR6), | 1029 | | forward_connected(L32, BFR7)}, seed1) | 1030 ------------------------------------------------------------------ 1031 BIFT entry in BFR4, BFR5: 1032 ------------------------------------------------------------------ 1033 | 0:8 | forward_connected(Lxx, BFR8) |xx differs on BFR4/BFR5| 1034 ------------------------------------------------------------------ 1036 BIFT entry in BFR6, BFR7: 1037 ------------------------------------------------------------------ 1038 | 0:8 | forward_connected(Lxx, BFR9) |xx differs on BFR6/BFR7| 1039 ------------------------------------------------------------------ 1041 BIFT entry in BFR8, BFR9: 1042 ------------------------------------------------------------------ 1043 | 0:9 | forward_connected(Lxx, BFR10) |xx differs on BFR8/BFR9| 1044 ------------------------------------------------------------------ 1046 Figure 13: Polarization Example 1048 Note that for the following discussion of ECMP, only the BIFT ECMP 1049 adjacencies on BFR1, BFR2, BFR3 are relevant. The re-use of BP 1050 across BFR in this example is further explained in Section 4.9 below. 1052 With the setup of ECMP in above topology, traffic would not be 1053 equally load-split. Instead, links L22 and L31 would see no traffic 1054 at all: BFR2 will only see traffic from BFR1 for which the ECMP hash 1055 in BFR1 selected the first adjacency in the list of 2 adjacencies 1056 given as parameters to the ECMP. It is link L11-to-BFR2. BFR2 1057 performs again ECMP with two adjacencies on that subset of traffic 1058 using the same seed1, and will therefore again select the first of 1059 its two adjacencies: L21-to-BFR4. And therefore L22 and BFR5 sees no 1060 traffic. Likewise for L31 and BFR6. 1062 This issue in BFR2/BFR3 is called polarization. It results from the 1063 re-use of the same hash function across multiple consecutive hops in 1064 topologies like these. To resolve this issue, the ECMP adjacency on 1065 BFR1 can be set up with a different seed2 than the ECMP adjacencies 1066 on BFR2/BFR3. BFR2/BFR3 can use the same hash because packets will 1067 not sequentially pass across both of them. Therefore, they can also 1068 use the same BP 0:7. 1070 Note that ECMP solutions outside of BIER often hide the seed by auto- 1071 selecting it from local entropy such as unique local or next-hop 1072 identifiers. The solutions choosen for BIER-TE to allow the BIER-TE 1073 Controller to explicitly set the seed maximizes the ability of the 1074 BIER-TE Controller to choose the seed, independent of such seed 1075 source that the BIER-TE Controller may not be able to control well, 1076 and even calculate optimized seeds for multi-hop cases. 1078 4.8. Routed adjacencies 1080 4.8.1. Reducing BitPositions 1082 Routed adjacencies can reduce the number of BitPositions required 1083 when the path steering requirement is not hop-by-hop explicit path 1084 selection, but loose-hop selection. Routed adjacencies can also 1085 allow to operate BIER-TE across intermediate hop routers that do not 1086 support BIER-TE. 1088 ............... 1089 ...BFR1--... ...--L1-- BFR2... 1090 ... .Routers. ...--L2--/ 1091 ...BFR4--... ...------ BFR3... 1092 ............... | 1093 LO 1094 Network Area 1 1096 Figure 14: Routed Adjacencies Example 1098 Assume the requirement in the above picture is to explicitly steer 1099 traffic flows that have arrived at BFR1 or BFR4 via a shortest path 1100 in the routing underlay "Network Area 1" to one of the following 1101 three next segments: (1) BFR2 via link L1, (2) BFR2 via link L2, (3) 1102 via BFR3. 1104 To enable this, both BFR1 and BFR4 are set up with a forward_routed 1105 adjacency BitPosition towards an address of BFR2 on link L1, another 1106 forward_routed BitPosition towards an address of BFR2 on link L2 and 1107 a third forward_routed Bitposition towards a node address LO of BFR3. 1109 4.8.2. Supporting nodes without BIER-TE 1111 Routed adjacencies also enable incremental deployment of BIER-TE. 1112 Only the nodes through which BIER-TE traffic needs to be steered - 1113 with or without replication - need to support BIER-TE. Where they 1114 are not directly connected to each other, forward_routed adjacencies 1115 are used to pass over non BIER-TE enabled nodes. 1117 4.9. Reuse of BitPositions (without DNR) 1119 BitPositions can be re-used across multiple BFR to minimize the 1120 number of BP needed. This happens when adjacencies on multiple BFR 1121 use the DNR flag as described above, but it can also be done for non- 1122 DNR adjacencies. This section only discussses this non-DNR case. 1124 Because BP are reset after passing a BFR with an adjacency for that 1125 BP, reuse of BP across multiple BFR does not introduce any problems 1126 with duplicates or loops that do not also exist when every adjacency 1127 has a unique BP: Instead of setting one BP in a BitString that is 1128 reused in N-adjacencies, one would get the same or worse results if 1129 each of these adjacencies had a unique BP and all of them where set 1130 in the BitString. Instead, based on the case, BPs can be reused 1131 without limitation, or they introduce fewer path steering choices, or 1132 they do not work. 1134 BP cannot be reused across two BFR that would need to be passed 1135 sequentially for some path: The first BFR will reset the BP, so those 1136 paths cannot be built. BP can be set across BFR that would (A) only 1137 occur across different paths or (B) across different branches of the 1138 same tree. 1140 An example of (A) was given in Figure 13, where BP 0:7, BP 0:8 and BP 1141 0:9 are each reused across multiple BFR because a single packet/path 1142 would never be able to reach more than one BFR sharing the same BP. 1144 Assume the example was changed: BFR1 has no ECMP adjacency for BP 1145 0:6, but instead BP 0:5 with forward_connected to BFR2 and BP 0:6 1146 with forward_connected to BFR3. Packets with both BP 0:5 and BP 0:6 1147 would now be able to reach both BFR2 and BFR3 and the still existing 1148 re-use of BP 0:7 between BFR2 and BFR3 is a case of (B) where reuse 1149 of BP is perfect because it does not limit the set of useful path 1150 choices: 1152 If instead of reusing BP 0:7, BFR3 used a separate BP 0:10 for its 1153 ECMP adjacency, no useful additional path steering options would be 1154 enabled. If duplicates at BFR10 where undesirable, this would be 1155 done by not setting BP 0:5 and BP 0:6 for the same packet. If the 1156 duplicates where desirable (e.g.: resilient transmission), the 1157 additional BP 0:10 would also not render additional value. 1159 Reuse may also save BPs in larger topologies. Consider the topology 1160 shown in Figure 17, but only the following explanations: A BFIR/ 1161 sender (e.g.: video headend) is attached to area 1, and area 2...6 1162 contain receivers/BFER. Assume each area had a distribution ring, 1163 each with two BPs to indicate the direction (as explained in before). 1164 These two BPs could be reused across the 5 areas. Packets would be 1165 replicated through other BPs to the desired subset of areas, and once 1166 a packet copy reaches the ring of the area, the two ring BPs come 1167 into play. This reuse is a case of (B), but it limits the topology 1168 choices: Packets can only flow around the same direction in the rings 1169 of all areas. This may or may not be acceptable based on the desired 1170 path steering options: If resilient transmission is the path 1171 engineering goal, then it is likely a good optimization, if the 1172 bandwidth of each ring was to be optimized separately, it would not 1173 be a good limitation. 1175 4.10. Summary of BP optimizations 1177 This section reviewed a range of techniques by which a BIER-TE 1178 Controller can create a BIER-TE topology in a way that minimizes the 1179 number of necessary BPs. 1181 Without any optimization, a BIER-TE Controller would attempt to map 1182 the network subnet topology 1:1 into the BIER-TE topology and every 1183 subnet adjacent neighbor requires a forward_connected BP and every 1184 BFER requires a local_decap BP. 1186 The optimizations described are then as follows: 1188 o P2p links require only one BP (Section 4.1). 1190 o All leaf-BFER can share a single local_decap BP (Section 4.3). 1192 o A LAN with N BFR needs at most N BP (one for each BFR). It only 1193 needs one BP for all those BFR tha are not redundanty connected to 1194 multiple LANs (Section 4.4). 1196 o A hub with p2p connections to multiple non-leaf-BFER spokes can 1197 share one BP to all spokes if traffic can be flooded to all 1198 spokes, e.g.: because of no bandwidth concerns or dense receiver 1199 sets (Section 4.5). 1201 o Rings of BFR can be built with just two BP (one for each 1202 direction) except for BFR with multiple ring connections - similar 1203 to LANs (Section 4.6). 1205 o ECMP adjacencies to N neighbors can replace N BP with 1 BP. 1206 Multihop ECMP can avoid polarization through different seeds of 1207 the ECMP algorithm (Section 4.7). 1209 o Routed adjacencies allow to "tunnel" across non-BIER-TE capable 1210 routers and across BIER-TE capable routers where no traffic- 1211 steering or replications are required (Section 4.8). 1213 o BP can generally be reused across nodes that do not need to be 1214 consecutive in paths, but depending on scenario, this may limit 1215 the feasible path steering options (Section 4.9). 1217 Note that the described list of optimizations is not exhaustive. 1218 Especially when the set of required path steering choices is limited 1219 and the set of possible subsets of BFER that should be able to 1220 receive traffic is limited, further optimizations of BP are possible. 1221 The hub & spoke optimization is a simple example of such traffic 1222 pattern dependent optimizations. 1224 5. Avoiding loops and duplicates 1226 5.1. Loops 1228 Whenever BIER-TE creates a copy of a packet, the BitString of that 1229 copy will have all BitPositions cleared that are associated with 1230 adjacencies on the BFR. This inhibits looping of packets. The only 1231 exception are adjacencies with DNR set. 1233 With DNR set, looping can happen. Consider in the ring picture that 1234 link L4 from BFR3 is plugged into the L1 interface of BFRa. This 1235 creates a loop where the rings clockwise BitPosition is never reset 1236 for copies of the packets traveling clockwise around the ring. 1238 To inhibit looping in the face of such physical misconfiguration, 1239 only forward_connected adjacencies are permitted to have DNR set, and 1240 the link layer port unique unicast destination address of the 1241 adjacency (e.g. MAC address) protects against closing the loop. 1242 Link layers without port unique link layer addresses should not be 1243 used with the DNR flag set. 1245 5.2. Duplicates 1247 Duplicates happen when the topology of the BitString is not a tree 1248 but redundantly connecting BFRs with each other. The BIER-TE 1249 Controller must therefore ensure to only create BitStrings that are 1250 trees in the topology. 1252 When links are incorrectly physically re-connected before the BIER-TE 1253 Controller updates BitStrings in BFIRs, duplicates can happen. Like 1254 loops, these can be inhibited by link layer addressing in 1255 forward_connected adjacencies. 1257 If interface or loopback addresses used in forward_routed adjacencies 1258 are moved from one BFR to another, duplicates can equally happen. 1259 Such re-addressing operations must be coordinated with the BIER-TE 1260 Controller. 1262 6. BIER-TE Forwarding Pseudocode 1264 The following simplified pseudocode for BIER-TE forwarding is using 1265 BIER forwarding pseudocode of [RFC8279], section 6.5 with the one 1266 modification necessary to support basic BIER-TE forwarding. Like the 1267 BIER pseudo forwarding code, for simplicity it does hide the details 1268 of the adjacency processing inside PacketSend() which can be 1269 forward_connected, forward_routed or local_decap. 1271 void ForwardBitMaskPacket_withTE (Packet) 1272 { 1273 SI=GetPacketSI(Packet); 1274 Offset=SI*BitStringLength; 1275 for (Index = GetFirstBitPosition(Packet->BitString); Index ; 1276 Index = GetNextBitPosition(Packet->BitString, Index)) { 1277 F-BM = BIFT[Index+Offset]->F-BM; 1278 if (!F-BM) continue; 1279 BFR-NBR = BIFT[Index+Offset]->BFR-NBR; 1280 PacketCopy = Copy(Packet); 1281 PacketCopy->BitString &= F-BM; [2] 1282 PacketSend(PacketCopy, BFR-NBR); 1283 // The following must not be done for BIER-TE: 1284 // Packet->BitString &= ~F-BM; [1] 1285 } 1286 } 1288 Figure 15: Simplified BIER-TE Forwarding Pseudocode 1290 The difference is that in BIER-TE, step [1] must not be performed, 1291 but is replaced with [2] (when the forwarding plane algorithm is 1292 implemented verbatim as shown above). 1294 In BIER, the F-BM of a BP has all BP set that are meant to be 1295 forwarded via the same neighbor. It is used to reset those BP in the 1296 packet after the first copy to this neighbor has been made to inhibit 1297 multiple copies to the same neighbor. 1299 In BIER-TE, the F-BM of a particular BP with an adjacency is the list 1300 of all BPs with an adjacency on this BFR except the particular BP 1301 itself if it has an adjacency with the DNR bit set. The F-BM is used 1302 to reset the F-BM BPs before creating copies. 1304 In BIER, the order of BPs impacts the result of forwarding because of 1305 [1]. In BIER-TE, forwarding is not impacted by the order of BPs. It 1306 is therefore possible to further optimize forwarding than in BIER. 1307 For example, BIER-TE forwarding can be parallelized such that a 1308 parallel instance (such as an egres linecard) can process any subset 1309 of BPs without any considerations for the other BPs - and without any 1310 prior, cross-BP shared processing. 1312 The above simplified pseudocode is elaborated further as follows: 1314 o This pseudocode eliminates per-bit F-BM, therefore reducing state 1315 by BitStringLength^2*SI and eliminating the need for per-packet- 1316 copy masking operation except for adjacencies with DNR flag set: 1318 * AdjacentBits[SI] are bits with a non-empty list of adjacencies. 1319 This can be computed whenever the BIER-TE Controller updates 1320 the adjacencies. 1322 * Only the AdjacentBits need to be examined in the loop for 1323 packet copies. 1325 * The packets BitString is masked with those AdjacentBits on 1326 ingress to avoid packets looping. 1328 o The code loops over the adjacencies because there may be more than 1329 one adjacency for a bit. 1331 o When an adjacency has the DNR bit, the bit is set in the packet 1332 copy (to save bits in rings for example). 1334 o The ECMP adjacency is shown. Its parameters are a 1335 ListOfAdjacencies from which one is picked. 1337 o The forward_local, forward_routed, local_decap adjacencies are 1338 shown with their parameters. 1340 void ForwardBitMaskPacket_withTE (Packet) 1341 { 1342 SI=GetPacketSI(Packet); 1343 Offset=SI*BitStringLength; 1344 AdjacentBitstring = Packet->BitString &= ~AdjacentBits[SI]; 1345 Packet->BitString &= AdjacentBits[SI]; 1346 for (Index = GetFirstBitPosition(AdjacentBits); Index ; 1347 Index = GetNextBitPosition(AdjacentBits, Index)) { 1348 foreach adjacency BIFT[Index+Offset] { 1349 if(adjacency == ECMP(ListOfAdjacencies, seed) ) { 1350 I = ECMP_hash(sizeof(ListOfAdjacencies), 1351 Packet->Entropy, seed); 1352 adjacency = ListOfAdjacencies[I]; 1353 } 1354 PacketCopy = Copy(Packet); 1355 switch(adjacency) { 1356 case forward_connected(interface,neighbor,DNR): 1357 if(DNR) 1358 PacketCopy->BitString |= 2<<(Index-1); 1359 SendToL2Unicast(PacketCopy,interface,neighbor); 1361 case forward_routed({VRF},neighbor): 1362 SendToL3(PacketCopy,{VRF,}l3-neighbor); 1364 case local_decap({VRF},neighbor): 1365 DecapBierHeader(PacketCopy); 1366 PassTo(PacketCopy,{VRF,}Packet->NextProto); 1367 } 1368 } 1369 } 1370 } 1372 Figure 16: BIER-TE Forwarding Pseudocode 1374 7. Managing SI, subdomains and BFR-ids 1376 When the number of bits required to represent the necessary hops in 1377 the topology and BFER exceeds the supported bitstring length, 1378 multiple SI and/or subdomains must be used. This section discusses 1379 how. 1381 BIER-TE forwarding does not require the concept of BFR-id, but 1382 routing underlay, flow overlay and BIER headers may. This section 1383 also discusses how BFR-ids can be assigned to BFIR/BFER for BIER-TE. 1385 7.1. Why SI and sub-domains 1387 For BIER and BIER-TE forwarding, the most important result of using 1388 multiple SI and/or subdomains is the same: Packets that need to be 1389 sent to BFER in different SI or subdomains require different BIER 1390 packets: each one with a bitstring for a different (SI,subdomain) 1391 combination. Each such bitstring uses one bitstring length sized SI 1392 block in the BIFT of the subdomain. We call this a BIFT:SI (block). 1394 For BIER and BIER-TE forwarding itself there is also no difference 1395 whether different SI and/or sub-domains are chosen, but SI and 1396 subdomain have different purposes in the BIER architecture shared by 1397 BIER-TE. This impacts how operators are managing them and how 1398 especially flow overlays will likely use them. 1400 By default, every possible BFIR/BFER in a BIER network would likely 1401 be given a BFR-id in subdomain 0 (unless there are > 64k BFIR/BFER). 1403 If there are different flow services (or service instances) requiring 1404 replication to different subsets of BFER, then it will likely not be 1405 possible to achieve the best replication efficiency for all of these 1406 service instances via subdomain 0. Ideal replication efficiency for 1407 N BFER exists in a subdomain if they are split over not more than 1408 ceiling(N/bitstring-length) SI. 1410 If service instances justify additional BIER:SI state in the network, 1411 additional subdomains will be used: BFIR/BFER are assigned BFIR-id in 1412 those subdomains and each service instance is configured to use the 1413 most appropriate subdomain. This results in improved replication 1414 efficiency for different services. 1416 Even if creation of subdomains and assignment of BFR-id to BFIR/BFER 1417 in those subdomains is automated, it is not expected that individual 1418 service instances can deal with BFER in different subdomains. A 1419 service instance may only support configuration of a single subdomain 1420 it should rely on. 1422 To be able to easily reuse (and modify as little as possible) 1423 existing BIER procedures including flow-overlay and routing underlay, 1424 when BIER-TE forwarding is added, we therefore reuse SI and subdomain 1425 logically in the same way as they are used in BIER: All necessary 1426 BFIR/BFER for a service use a single BIER-TE BIFT and are split 1427 across as many SI as necessary (see below). Different services may 1428 use different subdomains that primarily exist to provide more 1429 efficient replication (and for BIER-TE desirable path steering) for 1430 different subsets of BFIR/BFER. 1432 7.2. Bit assignment comparison BIER and BIER-TE 1434 In BIER, bitstrings only need to carry bits for BFER, which leads to 1435 the model that BFR-ids map 1:1 to each bit in a bitstring. 1437 In BIER-TE, bitstrings need to carry bits to indicate not only the 1438 receiving BFER but also the intermediate hops/links across which the 1439 packet must be sent. The maximum number of BFER that can be 1440 supported in a single bitstring or BIFT:SI depends on the number of 1441 bits necessary to represent the desired topology between them. 1443 "Desired" topology because it depends on the physical topology, and 1444 on the desire of the operator to allow for explicit path steeering 1445 across every single hop (which requires more bits), or reducing the 1446 number of required bits by exploiting optimizations such as unicast 1447 (forward_route), ECMP or flood (DNR) over "uninteresting" sub-parts 1448 of the topology - e.g. parts where different trees do not need to 1449 take different paths due to path steering reasons. 1451 The total number of bits to describe the topology vs. the BFER in a 1452 BIFT:SI can range widely based on the size of the topology and the 1453 amount of alternative paths in it. The higher the percentage, the 1454 higher the likelihood, that those topology bits are not just BIER-TE 1455 overhead without additional benefit, but instead that they will allow 1456 to express desirable path steering alternatives. 1458 7.3. Using BFR-id with BIER-TE 1460 Because there is no 1:1 mapping between bits in the bitstring and 1461 BFER, BIER-TE cannot simply rely on the BIER 1:1 mapping between bits 1462 in a bitstring and BFR-id. 1464 In BIER, automatic schemes could assign all possible BFR-ids 1465 sequentially to BFERs. This will not work in BIER-TE. In BIER-TE, 1466 the operator or BIER-TE Controller has to determine a BFR-id for each 1467 BFER in each required subdomain. The BFR-id may or may not have a 1468 relationship with a bit in the bitstring. Suggestions are detailed 1469 below. Once determined, the BFR-id can then be configured on the 1470 BFER and used by flow overlay, routing underlay and the BIER header 1471 almost the same as the BFR-id in BIER. 1473 The one exception are application/flow-overlays that automatically 1474 calculate the bitstring(s) of BIER packets by converting BFR-id to 1475 bits. In BIER-TE, this operation can be done in two ways: 1477 "Independent branches": For a given application or (set of) trees, 1478 the branches from a BFIR to every BFER are independent of the 1479 branches to any other BFER. For example, shortest part trees have 1480 independent branches. 1482 "Interdependent branches": When a BFER is added or deleted from a 1483 particular distribution tree, branches to other BFER still in the 1484 tree may need to change. Steiner tree are examples of dependent 1485 branch trees. 1487 If "independent branches" are sufficient, the BIER-TE Controller can 1488 provide to such applications for every BFR-id a SI:bitstring with the 1489 BIER-TE bits for the branch towards that BFER. The application can 1490 then independently calculate the SI:bitstring for all desired BFER by 1491 OR'ing their bitstrings. 1493 If "interdependent branches" are required, the application could call 1494 a BIER-TE Controller API with the list of required BFER-id and get 1495 the required bitstring back. Whenever the set of BFER-id changes, 1496 this is repeated. 1498 Note that in either case (unlike in BIER), the bits in BIER-TE may 1499 need to change upon link/node failure/recovery, network expansion and 1500 network resource consumption by other traffic as part of traffic 1501 engineering goals (e.g.: re-optimization of lower priority traffic 1502 flows). Interactions between such BFIR applications and the BIER-TE 1503 Controller do therefore need to support dynamic updates to the 1504 bitstrings. 1506 7.4. Assigning BFR-ids for BIER-TE 1508 For a non-leaf BFER, there is usually a single bit k for that BFER 1509 with a local_decap() adjacency on the BFER. The BFR-id for such a 1510 BFER is therefore most easily the one it would have in BIER: SI * 1511 bitstring-length + k. 1513 As explained earlier in the document, leaf BFERs do not need such a 1514 separate bit because the fact alone that the BIER-TE packet is 1515 forwarded to the leaf BFER indicates that the BFER should decapsulate 1516 it. Such a BFER will have one or more bits for the links leading 1517 only to it. The BFR-id could therefore most easily be the BFR-id 1518 derived from the lowest bit for those links. 1520 These two rules are only recommendations for the operator or BIER-TE 1521 Controller assigning the BFR-ids. Any allocation scheme can be used, 1522 the BFR-ids just need to be unique across BFRs in each subdomain. 1524 It is not currently determined if a single subdomain could or should 1525 be allowed to forward both BIER and BIER-TE packets. If this should 1526 be supported, there are two options: 1528 A. BIER and BIER-TE have different BFR-id in the same subdomain. 1529 This allows higher replication efficiency for BIER because their BFR- 1530 id can be assigned sequentially, while the bitstrings for BIER-TE 1531 will have also the additional bits for the topology. There is no 1532 relationship between a BFR BIER BFR-id and BIER-TE BFR-id. 1534 B. BIER and BIER-TE share the same BFR-id. The BFR-id are assigned 1535 as explained above for BIER-TE and simply reused for BIER. The 1536 replication efficiency for BIER will be as low as that for BIER-TE in 1537 this approach. Depending on topology, only the same 20%..80% of bits 1538 as possible for BIER-TE can be used for BIER. 1540 7.5. Example bit allocations 1542 7.5.1. With BIER 1544 Consider a network setup with a bitstring length of 256 for a network 1545 topology as shown in the picture below. The network has 6 areas, 1546 each with ca. 170 BFR, connecting via a core with some larger (core) 1547 BFR. To address all BFER with BIER, 4 SI are required. To send a 1548 BIER packet to all BFER in the network, 4 copies need to be sent by 1549 the BFIR. On the BFIR it does not make a difference how the BFR-id 1550 are allocated to BFER in the network, but for efficiency further down 1551 in the network it does make a difference. 1553 area1 area2 area3 1554 BFR1a BFR1b BFR2a BFR2b BFR3a BFR3b 1555 | \ / \ / | 1556 ................................ 1557 . Core . 1558 ................................ 1559 | / \ / \ | 1560 BFR4a BFR4b BFR5a BFR5b BFR6a BFR6b 1561 area4 area5 area6 1563 Figure 17: Scaling BIER-TE bits by reuse 1565 With random allocation of BFR-id to BFER, each receiving area would 1566 (most likely) have to receive all 4 copies of the BIER packet because 1567 there would be BFR-id for each of the 4 SI in each of the areas. 1568 Only further towards each BFER would this duplication subside - when 1569 each of the 4 trees runs out of branches. 1571 If BFR-id are allocated intelligently, then all the BFER in an area 1572 would be given BFR-id with as few as possible different SI. Each 1573 area would only have to forward one or two packets instead of 4. 1575 Given how networks can grow over time, replication efficiency in an 1576 area will also easily go down over time when BFR-id are network wide 1577 allocated sequentially over time. An area that initially only has 1578 BFR-id in one SI might end up with many SI over a longer period of 1579 growth. Allocating SIs to areas with initially sufficiently many 1580 spare bits for growths can help to alleviate this issue. Or renumber 1581 BFR-id after network expansion. In this example one may consider to 1582 use 6 SI and assign one to each area. 1584 This example shows that intelligent BFR-id allocation within at least 1585 subdomain 0 can even be helpful or even necessary in BIER. 1587 7.5.2. With BIER-TE 1589 In BIER-TE one needs to determine a subset of the physical topology 1590 and attached BFER so that the "desired" representation of this 1591 topology and the BFER fit into a single bitstring. This process 1592 needs to be repeated until the whole topology is covered. 1594 Once bits/SIs are assigned to topology and BFER, BFR-id is just a 1595 derived set of identifiers from the operator/BIER-TE Controller as 1596 explained above. 1598 Every time that different sub-topologies have overlap, bits need to 1599 be repeated across the bitstrings, increasing the overall amount of 1600 bits required across all bitstring/SIs. In the worst case, random 1601 subsets of BFER are assigned to different SI. This is much worse 1602 than in BIER because it not only reduces replication efficiency with 1603 the same number of overall bits, but even further - because more bits 1604 are required due to duplication of bits for topology across multiple 1605 SI. Intelligent BFER to SI assignment and selecting specific 1606 "desired" subtopologies can minimize this problem. 1608 To set up BIER-TE efficiently for above topology, the following bit 1609 allocation methods can be used. This method can easily be expanded 1610 to other, similarly structured larger topologies. 1612 Each area is allocated one or more SI depending on the number of 1613 future expected BFER and number of bits required for the topology in 1614 the area. In this example, 6 SI, one per area. 1616 In addition, we use 4 bits in each SI: bia, bib, bea, beb: bit 1617 ingress a, bit ingress b, bit egress a, bit egress b. These bits 1618 will be used to pass BIER packets from any BFIR via any combination 1619 of ingress area a/b BFR and egress area a/b BFR into a specific 1620 target area. These bits are then set up with the right 1621 forward_routed adjacencies on the BFIR and area edge BFR: 1623 On all BFIR in an area j, bia in each BIFT:SI is populated with the 1624 same forward_routed(BFRja), and bib with forward_routed(BFRjb). On 1625 all area edge BFR, bea in BIFT:SI=k is populated with 1626 forward_routed(BFRka) and beb in BIFT:SI=k with 1627 forward_routed(BFRkb). 1629 For BIER-TE forwarding of a packet to some subset of BFER across all 1630 areas, a BFIR would create at most 6 copies, with SI=1...SI=6, In 1631 each packet, the bits indicate bits for topology and BFER in that 1632 topology plus the four bits to indicate whether to pass this packet 1633 via the ingress area a or b border BFR and the egress area a or b 1634 border BFR, therefore allowing path steering for those two "unicast" 1635 legs: 1) BFIR to ingress are edge and 2) core to egress area edge. 1636 Replication only happens inside the egress areas. For BFER in the 1637 same area as in the BFIR, these four bits are not used. 1639 7.6. Summary 1641 BIER-TE can like BIER support multiple SI within a sub-domain to 1642 allow re-using the concept of BFR-id and therefore minimize BIER-TE 1643 specific functions in underlay routing, flow overlay methods and BIER 1644 headers. 1646 The number of BFIR/BFER possible in a subdomain is smaller than in 1647 BIER because BIER-TE uses additional bits for topology. 1649 Subdomains can in BIER-TE be used like in BIER to create more 1650 efficient replication to known subsets of BFER. 1652 Assigning bits for BFER intelligently into the right SI is more 1653 important in BIER-TE than in BIER because of replication efficiency 1654 and overall amount of bits required. 1656 8. BIER-TE and Segment Routing 1658 SR aims to enable lightweight path steering via loose source routing. 1659 Compared to its more heavy-weight predecessor RSVP-TE, SR does for 1660 example not require per-path signaling to each of these hops. 1662 BIER-TE supports the same design philosophy for multicast. Like in 1663 SR, it relies on source-routing - via the definition of a BitString. 1664 Like SR, it only requires to consider the "hops" on which either 1665 replication has to happen, or across which the traffic should be 1666 steered (even without replication). Any other hops can be skipped 1667 via the use of routed adjacencies. 1669 BIER-TE BitPosition (BP) can be understood as the BIER-TE equivalent 1670 of "forwarding segments" in SR, but they have a different scope than 1671 SR forwarding segments. Whereas forwarding segments in SR are global 1672 or local, BPs in BIER-TE have a scope that is the group of BFR(s) 1673 that have adjacencies for this BP in their BIFT. This can be called 1674 "adjacency" scoped forwarding segments. 1676 Adjacency scope could be global, but then every BFR would need an 1677 adjacency for this BP, for example a forward_routed adjacency with 1678 encapsulation to the global SR SID of the destination. Such a BP 1679 would always result in ingress replication though. The first BFR 1680 encountering this BP would directly replicate to it. Only by using 1681 non-global adjacency scope for BPs can traffic be steered and 1682 replicated on non-ingress BFR. 1684 SR can naturally be combined with BIER-TE and help to optimize it. 1685 For example, instead of defining BitPositions for non-replicating 1686 hops, it is equally possible to use segment routing encapsulations 1687 (eg: MPLS label stacks) for the encapsulation of "forward_routed" 1688 adjacencies. 1690 Note that BIER itself can also be seen to be similar to SR. BIER BPs 1691 act as global destination Node-SIDs and the BIER bitstring is simply 1692 a highly optimized mechanism to indicate multiple such SIDS and let 1693 the network take care of effectively replicating the packet hop-by- 1694 hop to each destination Node-SID. What BIER does not allow is to 1695 indicate intermediate hops, or terms of SR the ability to indicate a 1696 sequence of SID to reach the destination. This is what BIER-TE and 1697 its adjacency scoped BP enables. 1699 Both BIER and BIER-TE allow BFIR to "opportunistically" copy packets 1700 to a set of desired BFER on a packet-by-packet basis. In BIER, this 1701 is done by OR'ing the BP for the desired BFER. In BIER-TE this can 1702 be done by OR'ing for each desired BFER a bitstring using the 1703 "independent branches" approach described in Section 7.3 and 1704 therefore also indicating the engineered path towards each desired 1705 BFER. This is the approach that 1706 [I-D.ietf-bier-multicast-http-response] relies on. 1708 9. Security Considerations 1710 The security considerations are the same as for BIER with the 1711 following differences: 1713 BFR-ids and BFR-prefixes are not used in BIER-TE, nor are procedures 1714 for their distribution, so these are not attack vectors against BIER- 1715 TE. 1717 10. IANA Considerations 1719 This document requests no action by IANA. 1721 11. Acknowledgements 1723 The authors would like to thank Greg Shepherd, Ijsbrand Wijnands, 1724 Neale Ranns, Dirk Trossen, Sandy Zheng, Lou Berger and Jeffrey Zhang 1725 for their reviews and suggestions. 1727 12. Change log [RFC Editor: Please remove] 1729 draft-ietf-bier-te-arch: 1731 07: Further reworking text for Lou. 1733 Renamed BIER-PE to BIER-TE standing for "Tree Engineering" after 1734 votes from BIER WG. 1736 Removed section 1.1 (introduced by version 06) because not 1737 considered necessary in this doc by Lou (for framework doc). 1739 Added [RFC editor pls. remove] Section to explain name change to 1740 future reviewers. 1742 06: Concern by Lou Berger re. BIER-TE as full traffic engineering 1743 solution. 1745 Changed title "Traffic Engineering" to "Path Engineering" 1747 Added intro section of relationship BIER-PE to traffic 1748 engineering. 1750 Changed "traffic engineering" term in text" to "path engineering", 1751 where appropriate 1753 Other: 1755 Shortened "BIER-TE Controller Host" to "BIER-TE Controller". 1756 Fixed up all instances of controller to do this. 1758 05: Review Jeffrey Zhang. 1760 Part 2: 1762 4.3 added note about leaf-BFER being also a propery of routing 1763 setup. 1765 4.7 Added missing details from example to avoid confusion with 1766 routed adjacencies, also compressed explanatory text and better 1767 justification why seed is explicitly configured by controller. 1769 4.9 added section discussing generic reuse of BP methods. 1771 4.10 added section summarizing BP optimizations of section 4. 1773 6. Rewrote/compressed explanation of comparison BIER/BIER-TE 1774 forwarding difference. Explained benefit of BIER-TE per-BP 1775 forwarding being independent of forwarding for other BPs. 1777 Part 1: 1779 Explicitly ue forwarded_connected adjcency in ECMP adjcency 1780 examples to avoid confusion. 1782 4.3 Add picture as example for leav vs. non-leaf BFR in topology. 1783 Improved description. 1785 4.5 Exampe for traffic that can be broadcast -> for single BP in 1786 hub&spoke. 1788 4.8.1 Simplified example picture for routed adjacency, explanatory 1789 text. 1791 Review from Dirk Trossen: 1793 Fixed up explanation of ICC paper vs. bloom filter. 1795 04: spell check run. 1797 Addded remaining fixes for Sandys (Zhang Zheng) review: 1799 4.7 Enhance ECMP explanations: 1801 example ECMP algorithm, highlight that doc does not standardize 1802 ECMP algorithm. 1804 Review from Dirk Trossen: 1806 1. Added mentioning of prior work for traffic engineered paths 1807 with bloom filters. 1809 2. Changed title from layers to components and added "BIER-TE 1810 control plane" to "BIER-TE Controller" to make it clearer, what it 1811 does. 1813 2.2.3. Added reference to I-D.ietf-bier-multicast-http-response 1814 as an example solution. 1816 2.3. clarified sentence about resetting BPs before sending copies 1817 (also forgot to mention DNR here). 1819 3.4. Added text saying this section will be removed unless IESG 1820 review finds enough redeeming value in this example given how -03 1821 introduced section 1.1 with basic examples. 1823 7.2. Removed explicit numbers 20%/80% for number of topology bits 1824 in BIER-TE, replaced with more vague (high/low) description, 1825 because we do not have good reference material Added text saying 1826 this section will be removed unless IESG review finds enough 1827 redeeming value in this example given how -03 introduced section 1828 1.1 with basic examples. 1830 many typos fixed. Thanks a lot. 1832 03: Last call textual changes by authors to improve readability: 1834 removed Wolfgang Braun as co-authors (as requested). 1836 Improved abstract to be more explanatory. Removed mentioning of 1837 FRR (not concluded on so far). 1839 Added new text into Introduction section because the text was too 1840 difficult to jump into (too many forward pointers). This 1841 primarily consists of examples and the early introduction of the 1842 BIER-TE Topology concept enabled by these examples. 1844 Amended comparison to SR. 1846 Changed syntax from [VRF] to {VRF} to indicate its optional and to 1847 make idnits happy. 1849 Split references into normative / informative, added references. 1851 02: Refresh after IETF104 discussion: changed intended status back 1852 to standard. Reasoning: 1854 Tighter review of standards document == ensures arch will be 1855 better prepared for possible adoption by other WGs (e.g. DetNet) 1856 or std. bodies. 1858 Requirement against the degree of existing implementations is self 1859 defined by the WG. BIER WG seems to think it is not necessary to 1860 apply multiple interoperating implementations against an 1861 architecture level document at this time to make it qualify to go 1862 to standards track. Also, the levels of support introduced in -01 1863 rev. should allow all BIER forwarding engines to also be able to 1864 support the base level BIER-TE forwarding. 1866 01: Added note comparing BIER and SR to also hopefully clarify 1867 BIER-TE vs. BIER comparison re. SR. 1869 - added requirements section mandating only most basic BIER-TE 1870 forwarding features as MUST. 1872 - reworked comparison with BIER forwarding section to only 1873 summarize and point to pseudocode section. 1875 - reworked pseudocode section to have one pseudocode that mirrors 1876 the BIER forwarding pseudocode to make comparison easier and a 1877 second pseudocode that shows the complete set of BIER-TE 1878 forwarding options and simplification/optimization possible vs. 1879 BIER forwarding. Removed MyBitsOfInterest (was pure 1880 optimization). 1882 - Added captions to pictures. 1884 - Part of review feedback from Sandy (Zhang Zheng) integrated. 1886 00: Changed target state to experimental (WG conclusion), updated 1887 references, mod auth association. 1889 - Source now on http://www.github.com/toerless/bier-te-arch 1891 - Please open issues on the github for change/improvement requests 1892 to the document - in addition to posting them on the list 1893 (bier@ietf.). Thanks!. 1895 draft-eckert-bier-te-arch: 1897 06: Added overview of forwarding differences between BIER, BIER- 1898 TE. 1900 05: Author affiliation change only. 1902 04: Added comparison to Live-Live and BFIR to FRR section 1903 (Eckert). 1905 04: Removed FRR content into the new FRR draft [I-D.eckert-bier- 1906 te-frr] (Braun). 1908 - Linked FRR information to new draft in Overview/Introduction 1909 - Removed BTAFT/FRR from "Changes in the network topology" 1911 - Linked new draft in "Link/Node Failures and Recovery" 1913 - Removed FRR from "The BIER-TE Forwarding Layer" 1915 - Moved FRR section to new draft 1917 - Moved FRR parts of Pseudocode into new draft 1919 - Left only non FRR parts 1921 - removed FrrUpDown(..) and //FRR operations in 1922 ForwardBierTePacket(..) 1924 - New draft contains FrrUpDown(..) and ForwardBierTePacket(Packet) 1925 from bier-arch-03 1927 - Moved "BIER-TE and existing FRR to new draft 1929 - Moved "BIER-TE and Segment Routing" section one level up 1931 - Thus, removed "Further considerations" that only contained this 1932 section 1934 - Added Changes for version 04 1936 03: Updated the FRR section. Added examples for FRR key concepts. 1937 Added BIER-in-BIER tunneling as option for tunnels in backup 1938 paths. BIFT structure is expanded and contains an additional 1939 match field to support full node protection with BIER-TE FRR. 1941 03: Updated FRR section. Explanation how BIER-in-BIER 1942 encapsulation provides P2MP protection for node failures even 1943 though the routing underlay does not provide P2MP. 1945 02: Changed the definition of BIFT to be more inline with BIER. 1946 In revs. up to -01, the idea was that a BIFT has only entries for 1947 a single bitstring, and every SI and subdomain would be a separate 1948 BIFT. In BIER, each BIFT covers all SI. This is now also how we 1949 define it in BIER-TE. 1951 02: Added Section 7 to explain the use of SI, subdomains and BFR- 1952 id in BIER-TE and to give an example how to efficiently assign 1953 bits for a large topology requiring multiple SI. 1955 02: Added further detailed for rings - how to support input from 1956 all ring nodes. 1958 01: Fixed BFIR -> BFER for section 4.3. 1960 01: Added explanation of SI, difference to BIER ECMP, 1961 consideration for Segment Routing, unicast FRR, considerations for 1962 encapsulation, explanations of BIER-TE Controller and CLI. 1964 00: Initial version. 1966 13. References 1968 13.1. Normative References 1970 [RFC8279] Wijnands, IJ., Ed., Rosen, E., Ed., Dolganow, A., 1971 Przygienda, T., and S. Aldrin, "Multicast Using Bit Index 1972 Explicit Replication (BIER)", RFC 8279, 1973 DOI 10.17487/RFC8279, November 2017, 1974 . 1976 [RFC8296] Wijnands, IJ., Ed., Rosen, E., Ed., Dolganow, A., 1977 Tantsura, J., Aldrin, S., and I. Meilik, "Encapsulation 1978 for Bit Index Explicit Replication (BIER) in MPLS and Non- 1979 MPLS Networks", RFC 8296, DOI 10.17487/RFC8296, January 1980 2018, . 1982 13.2. Informative References 1984 [I-D.ietf-bier-multicast-http-response] 1985 Trossen, D., Rahman, A., Wang, C., and T. Eckert, 1986 "Applicability of BIER Multicast Overlay for Adaptive 1987 Streaming Services", draft-ietf-bier-multicast-http- 1988 response-03 (work in progress), February 2020. 1990 [I-D.ietf-roll-ccast] 1991 Bergmann, O., Bormann, C., Gerdes, S., and H. Chen, 1992 "Constrained-Cast: Source-Routed Multicast for RPL", 1993 draft-ietf-roll-ccast-01 (work in progress), October 2017. 1995 [ICC] Reed, M., Al-Naday, M., Thomos, N., Trossen, D., 1996 Petropoulos, G., and S. Spirou, "Stateless multicast 1997 switching in software defined networks", IEEE 1998 International Conference on Communications (ICC), Kuala 1999 Lumpur, Malaysia, 2016, May 2016, 2000 . 2002 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 2003 Requirement Levels", BCP 14, RFC 2119, 2004 DOI 10.17487/RFC2119, March 1997, 2005 . 2007 Authors' Addresses 2009 Toerless Eckert (editor) 2010 Futurewei Technologies Inc. 2011 2330 Central Expy 2012 Santa Clara 95050 2013 USA 2015 Email: tte+ietf@cs.fau.de 2017 Gregory Cauchie 2018 Bouygues Telecom 2020 Email: GCAUCHIE@bouyguestelecom.fr 2022 Michael Menth 2023 University of Tuebingen 2025 Email: menth@uni-tuebingen.de