idnits 2.17.1 draft-ietf-dmm-srv6-mobile-uplane-17.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 (12 October 2021) is 917 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: 'SL' is mentioned on line 174, but not defined == Missing Reference: 'N11' is mentioned on line 246, but not defined == Missing Reference: 'N2' is mentioned on line 247, but not defined == Missing Reference: 'N4' is mentioned on line 251, but not defined == Missing Reference: 'N3' is mentioned on line 698, but not defined == Missing Reference: 'N9' is mentioned on line 537, but not defined == Missing Reference: 'N6' is mentioned on line 699, but not defined -- Looks like a reference, but probably isn't: '0' on line 1060 == Missing Reference: 'NET-PGM' is mentioned on line 1111, but not defined == Outdated reference: A later version (-22) exists of draft-ietf-spring-segment-routing-policy-13 == Outdated reference: A later version (-26) exists of draft-ietf-lsr-flex-algo-17 == Outdated reference: A later version (-09) exists of draft-ietf-spring-sr-service-programming-05 == Outdated reference: A later version (-07) exists of draft-kohno-dmm-srv6mob-arch-04 == Outdated reference: A later version (-15) exists of draft-matsushima-spring-srv6-deployment-status-11 == Outdated reference: A later version (-05) exists of draft-murakami-dmm-user-plane-message-encoding-04 Summary: 0 errors (**), 0 flaws (~~), 15 warnings (==), 2 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 DMM Working Group S. Matsushima, Ed. 3 Internet-Draft SoftBank 4 Intended status: Standards Track C. Filsfils 5 Expires: 15 April 2022 M. Kohno 6 P. Camarillo, Ed. 7 Cisco Systems, Inc. 8 D. Voyer 9 Bell Canada 10 C.E. Perkins 11 Lupin Lodge 12 12 October 2021 14 Segment Routing IPv6 for Mobile User Plane 15 draft-ietf-dmm-srv6-mobile-uplane-17 17 Abstract 19 This document specifies the applicability of SRv6 (Segment Routing 20 IPv6) to the user-plane of mobile networks. The network programming 21 nature of SRv6 accomplishes mobile user-plane functions in a simple 22 manner. The statelessness of SRv6 and its ability to control both 23 service layer path and underlying transport can be beneficial to the 24 mobile user-plane, providing flexibility, end-to-end network slicing, 25 and SLA control for various applications. 27 Status of This Memo 29 This Internet-Draft is submitted in full conformance with the 30 provisions of BCP 78 and BCP 79. 32 Internet-Drafts are working documents of the Internet Engineering 33 Task Force (IETF). Note that other groups may also distribute 34 working documents as Internet-Drafts. The list of current Internet- 35 Drafts is at https://datatracker.ietf.org/drafts/current/. 37 Internet-Drafts are draft documents valid for a maximum of six months 38 and may be updated, replaced, or obsoleted by other documents at any 39 time. It is inappropriate to use Internet-Drafts as reference 40 material or to cite them other than as "work in progress." 42 This Internet-Draft will expire on 15 April 2022. 44 Copyright Notice 46 Copyright (c) 2021 IETF Trust and the persons identified as the 47 document authors. All rights reserved. 49 This document is subject to BCP 78 and the IETF Trust's Legal 50 Provisions Relating to IETF Documents (https://trustee.ietf.org/ 51 license-info) in effect on the date of publication of this document. 52 Please review these documents carefully, as they describe your rights 53 and restrictions with respect to this document. Code Components 54 extracted from this document must include Simplified BSD License text 55 as described in Section 4.e of the Trust Legal Provisions and are 56 provided without warranty as described in the Simplified BSD License. 58 Table of Contents 60 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 3 61 2. Conventions and Terminology . . . . . . . . . . . . . . . . . 3 62 2.1. Terminology . . . . . . . . . . . . . . . . . . . . . . . 3 63 2.2. Conventions . . . . . . . . . . . . . . . . . . . . . . . 4 64 2.3. Predefined SRv6 Endpoint Behaviors . . . . . . . . . . . 4 65 3. Motivation . . . . . . . . . . . . . . . . . . . . . . . . . 5 66 4. 3GPP Reference Architecture . . . . . . . . . . . . . . . . . 5 67 5. User-plane behaviors . . . . . . . . . . . . . . . . . . . . 6 68 5.1. Traditional mode . . . . . . . . . . . . . . . . . . . . 7 69 5.1.1. Packet flow - Uplink . . . . . . . . . . . . . . . . 8 70 5.1.2. Packet flow - Downlink . . . . . . . . . . . . . . . 9 71 5.2. Enhanced Mode . . . . . . . . . . . . . . . . . . . . . . 9 72 5.2.1. Packet flow - Uplink . . . . . . . . . . . . . . . . 10 73 5.2.2. Packet flow - Downlink . . . . . . . . . . . . . . . 11 74 5.2.3. Scalability . . . . . . . . . . . . . . . . . . . . . 11 75 5.3. Enhanced mode with unchanged gNB GTP behavior . . . . . . 12 76 5.3.1. Interworking with IPv6 GTP . . . . . . . . . . . . . 12 77 5.3.2. Interworking with IPv4 GTP . . . . . . . . . . . . . 15 78 5.3.3. Extensions to the interworking mechanisms . . . . . . 17 79 5.4. SRv6 Drop-in Interworking . . . . . . . . . . . . . . . . 18 80 6. SRv6 Segment Endpoint Mobility Behaviors . . . . . . . . . . 19 81 6.1. Args.Mob.Session . . . . . . . . . . . . . . . . . . . . 19 82 6.2. End.MAP . . . . . . . . . . . . . . . . . . . . . . . . . 20 83 6.3. End.M.GTP6.D . . . . . . . . . . . . . . . . . . . . . . 20 84 6.4. End.M.GTP6.D.Di . . . . . . . . . . . . . . . . . . . . . 22 85 6.5. End.M.GTP6.E . . . . . . . . . . . . . . . . . . . . . . 23 86 6.6. End.M.GTP4.E . . . . . . . . . . . . . . . . . . . . . . 24 87 6.7. H.M.GTP4.D . . . . . . . . . . . . . . . . . . . . . . . 25 88 6.8. End.Limit: Rate Limiting behavior . . . . . . . . . . . . 26 89 7. SRv6 supported 3GPP PDU session types . . . . . . . . . . . . 26 90 8. Network Slicing Considerations . . . . . . . . . . . . . . . 26 91 9. Control Plane Considerations . . . . . . . . . . . . . . . . 27 92 10. Security Considerations . . . . . . . . . . . . . . . . . . . 27 93 11. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 27 94 12. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 28 95 13. Contributors . . . . . . . . . . . . . . . . . . . . . . . . 28 96 14. References . . . . . . . . . . . . . . . . . . . . . . . . . 28 97 14.1. Normative References . . . . . . . . . . . . . . . . . . 28 98 14.2. Informative References . . . . . . . . . . . . . . . . . 29 99 Appendix A. Implementations . . . . . . . . . . . . . . . . . . 31 100 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 31 102 1. Introduction 104 In mobile networks, mobility systems provide connectivity over a 105 wireless link to stationary and non-stationary nodes. The user-plane 106 establishes a tunnel between the mobile node and its anchor node over 107 IP-based backhaul and core networks. 109 This document specifies the applicability of SRv6 (Segment Routing 110 IPv6) to mobile networks. 112 Segment Routing [RFC8402] is a source routing architecture: a node 113 steers a packet through an ordered list of instructions called 114 "segments". A segment can represent any instruction, topological or 115 service based. 117 SRv6 applied to mobile networks enables a source-routing based mobile 118 architecture, where operators can explicitly indicate a route for the 119 packets to and from the mobile node. The SRv6 Endpoint nodes serve 120 as mobile user-plane anchors. 122 2. Conventions and Terminology 124 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 125 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 126 document are to be interpreted as described in [RFC2119]. 128 2.1. Terminology 130 * CNF: Cloud-native Network Function 131 * NFV: Network Function Virtualization 132 * PDU: Packet Data Unit 133 * PDU Session: Context of a UE connects to a mobile network. 134 * UE: User Equipment 135 * UPF: User Plane Function 136 * VNF: Virtual Network Function (including CNFs) 138 The following terms used within this document are defined in 139 [RFC8402]: Segment Routing, SR Domain, Segment ID (SID), SRv6, SRv6 140 SID, Active Segment, SR Policy, Prefix SID, Adjacency SID and Binding 141 SID. 143 The following terms used within this document are defined in 144 [RFC8754]: SRH, SR Source Node, Transit Node, SR Segment Endpoint 145 Node and Reduced SRH. 147 The following terms used within this document are defined in 148 [RFC8986]: NH, SL, FIB, SA, DA, SRv6 SID behavior, SRv6 Segment 149 Endpoint Behavior. 151 2.2. Conventions 153 An SR Policy is resolved to a SID list. A SID list is represented as 154 where S1 is the first SID to visit, S2 is the second SID 155 to visit, and S3 is the last SID to visit along the SR path. 157 (SA,DA) (S3, S2, S1; SL) represents an IPv6 packet with: 159 * Source Address is SA, Destination Address is DA, and next-header 160 is SRH 161 * SRH with SID list with Segments Left = SL 162 * Note the difference between the <> and () symbols: 163 represents a SID list where S1 is the first SID and S3 is the last 164 SID to traverse. (S3, S2, S1; SL) represents the same SID list 165 but encoded in the SRH format where the rightmost SID in the SRH 166 is the first SID and the leftmost SID in the SRH is the last SID. 167 When referring to an SR policy in a high-level use-case, it is 168 simpler to use the notation. When referring to an 169 illustration of the detailed packet behavior, the (S3, S2, S1; SL) 170 notation is more convenient. 171 * The payload of the packet is omitted. 173 SRH[n]: A shorter representation of Segment List[n], as defined in 174 [RFC8754]. SRH[SL] can be different from the DA of the IPv6 header. 176 * gNB::1 is an IPv6 address (SID) assigned to the gNB. 177 * U1::1 is an IPv6 address (SID) assigned to UPF1. 178 * U2::1 is an IPv6 address (SID) assigned to UPF2. 179 * U2:: is some other IPv6 address (SID) assigned to UPF2. 181 2.3. Predefined SRv6 Endpoint Behaviors 183 The following SRv6 Endpoint Behaviors are defined in [RFC8986]. 185 * End.DT4: Decapsulation and Specific IPv4 Table Lookup 186 * End.DT6: Decapsulation and Specific IPv6 Table Lookup 187 * End.DT46: Decapsulation and Specific IP Table Lookup 188 * End.DX4: Decapsulation and IPv4 Cross-Connect 189 * End.DX6: Decapsulation and IPv6 Cross-Connect 190 * End.DX2: Decapsulation and L2 Cross-Connect 191 * End.T: Endpoint with specific IPv6 Table Lookup 193 This document defines new SRv6 Segment Endpoint Behaviors in 194 Section 6. 196 3. Motivation 198 Mobile networks are becoming more challenging to operate. On one 199 hand, traffic is constantly growing, and latency requirements are 200 tighter; on the other-hand, there are new use-cases like distributed 201 NFVi that are also challenging network operations. 203 The current architecture of mobile networks does not take into 204 account the underlying transport. The user-plane is rigidly 205 fragmented into radio access, core and service networks, connected by 206 tunneling according to user-plane roles such as access and anchor 207 nodes. These factors have made it difficult for the operator to 208 optimize and operate the data-path. 210 In the meantime, applications have shifted to use IPv6, and network 211 operators have started adopting IPv6 as their IP transport. SRv6, 212 the IPv6 dataplane instantiation of Segment Routing [RFC8402], 213 integrates both the application data-path and the underlying 214 transport layer into a single protocol, allowing operators to 215 optimize the network in a simplified manner and removing forwarding 216 state from the network. It is also suitable for virtualized 217 environments, like VNF/CNF to VNF/CNF networking. SRv6 has been 218 deployed in dozens of networks 219 [I-D.matsushima-spring-srv6-deployment-status]. 221 SRv6 defines the network-programming concept [RFC8986]. Applied to 222 mobility, SRv6 can provide the user-plane behaviors needed for 223 mobility management. SRv6 takes advantage of the underlying 224 transport awareness and flexibility together with the ability to also 225 include services to optimize the end-to-end mobile dataplane. 227 The use-cases for SRv6 mobility are discussed in 228 [I-D.camarilloelmalky-springdmm-srv6-mob-usecases], and the 229 architetural benefits are discussed in [I-D.kohno-dmm-srv6mob-arch]. 231 4. 3GPP Reference Architecture 233 This section presents a reference architecture and possible 234 deployment scenarios. 236 Figure 1 shows a reference diagram from the 5G packet core 237 architecture [TS.23501]. 239 The user plane described in this document does not depend on any 240 specific architecture. The 5G packet core architecture as shown is 241 based on the latest 3GPP standards at the time of writing this draft. 243 +-----+ 244 | AMF | 245 /+-----+ 246 / | [N11] 247 [N2] / +-----+ 248 +------/ | SMF | 249 / +-----+ 250 / / \ 251 / / \ [N4] 252 / / \ ________ 253 / / \ / \ 254 +--+ +-----+ [N3] +------+ [N9] +------+ [N6] / \ 255 |UE|------| gNB |------| UPF1 |--------| UPF2 |--------- \ DN / 256 +--+ +-----+ +------+ +------+ \________/ 258 Figure 1: 3GPP 5G Reference Architecture 260 * UE: User Endpoint 261 * gNB: gNodeB with N3 interface towards packet core (and N2 for 262 control plane) 263 * UPF1: UPF with Interfaces N3 and N9 (and N4 for control plane) 264 * UPF2: UPF with Interfaces N9 and N6 (and N4 for control plane) 265 * SMF: Session Management Function 266 * AMF: Access and Mobility Management Function 267 * DN: Data Network e.g. operator services, Internet access 269 This reference diagram does not depict a UPF that is only connected 270 to N9 interfaces, although the mechanisms defined in this document 271 also work in such case. 273 Each session from a UE gets assigned to a UPF. Sometimes multiple 274 UPFs may be used, providing richer service functions. A UE gets its 275 IP address from the DHCP block of its UPF. The UPF advertises that 276 IP address block toward the Internet, ensuring that return traffic is 277 routed to the right UPF. 279 5. User-plane behaviors 281 This section introduces an SRv6 based mobile user-plane. 283 In order to simplify the adoption of SRv6, we present two different 284 "modes" that vary with respect to the use of SRv6. The first one is 285 the "Traditional mode", which inherits the current 3GPP mobile 286 architecture. In this mode GTP-U protocol [TS.29281] is replaced by 287 SRv6, however the N3, N9 and N6 interfaces are still point-to-point 288 interfaces with no intermediate waypoints as in the current mobile 289 network architecture. 291 The second mode is the "Enhanced mode". This is an evolution from 292 the "Traditional mode". In this mode the N3, N9 or N6 interfaces 293 have intermediate waypoints -SIDs- that are used for Traffic 294 Engineering or VNF purposes transparent to 3GPP functionalities. 295 This results in optimal end-to-end policies across the mobile network 296 with transport and services awareness. 298 In both, the Traditional and the Enhanced modes, we assume that the 299 gNB as well as the UPFs are SR-aware (N3, N9 and -potentially- N6 300 interfaces are SRv6). 302 In addition to those two modes, we introduce two mechanisms for 303 interworking with legacy access networks (those where the N3 304 interface is unmodified). In this document we introduce them as a 305 variant to the Enhanced mode, however they are equally applicable to 306 the Traditional mode. 308 One of these mechanisms is designed to interwork with legacy gNBs 309 using GTP/IPv4. The second mechanism is designed to interwork with 310 legacy gNBs using GTP/IPv6. 312 This document uses SRv6 Segment Endpoint Behaviors defined in 313 [RFC8986] as well as new SRv6 Segment Endpoint Behaviors designed for 314 the mobile user plane that are defined in this document in Section 6. 316 5.1. Traditional mode 318 In the traditional mode, the existing mobile UPFs remain unchanged 319 with the sole exception of the use of SRv6 as the data plane instead 320 of GTP-U. There is no impact to the rest of the mobile system. 322 In existing 3GPP mobile networks, a PDU Session is mapped 1-for-1 323 with a specific GTP tunnel (TEID). This 1-for-1 mapping is mirrored 324 here to replace GTP encapsulation with the SRv6 encapsulation, while 325 not changing anything else. There will be a unique SRv6 SID 326 associated with each PDU Session, and the SID list only contains a 327 single SID. 329 The traditional mode minimizes the changes required to the mobile 330 system; hence it is a good starting point for forming a common 331 ground. 333 The gNB/UPF control-plane (N2/N4 interface) is unchanged, 334 specifically a single IPv6 address is provided to the gNB. The same 335 control plane signalling is used, and the gNB/UPF decides to use SRv6 336 based on signaled GTP-U parameters per local policy. The only 337 information from the GTP-U parameters used for the SRv6 policy is the 338 TEID, QFI, and the IPv6 Destination Address. 340 Our example topology is shown in Figure 2. The gNB and the UPFs are 341 SR-aware. In the descriptions of the uplink and downlink packet 342 flow, A is an IPv6 address of the UE, and Z is an IPv6 address 343 reachable within the Data Network DN. A new SRv6 Endpoint Behavior, 344 End.MAP, defined in Section 6.2, is used. 346 ________ 347 SRv6 SRv6 / \ 348 +--+ +-----+ [N3] +------+ [N9] +------+ [N6] / \ 349 |UE|------| gNB |------| UPF1 |--------| UPF2 |--------- \ DN / 350 +--+ +-----+ +------+ +------+ \________/ 351 SRv6 node SRv6 node SRv6 node 353 Figure 2: Traditional mode - example topology 355 5.1.1. Packet flow - Uplink 357 The uplink packet flow is as follows: 359 UE_out : (A,Z) 360 gNB_out : (gNB, U1::1) (A,Z) -> H.Encaps.Red 361 UPF1_out: (gNB, U2::1) (A,Z) -> End.MAP 362 UPF2_out: (A,Z) -> End.DT4 or End.DT6 364 When the UE packet arrives at the gNB, the gNB performs a 365 H.Encaps.Red operation. Since there is only one SID, there is no 366 need to push an SRH. gNB only adds an outer IPv6 header with IPv6 DA 367 U1::1. gNB obtains the SID U1::1 from the existing control plane (N2 368 interface). U1::1 represents an anchoring SID specific for that 369 session at UPF1. 371 When the packet arrives at UPF1, the SID U1::1 is associated with the 372 End.MAP SRv6 Endpoint Behavior. End.MAP replaces U1::1 by U2::1, 373 that belongs to the next UPF (U2). 375 When the packet arrives at UPF2, the SID U2::1 corresponds to an 376 End.DT4/End.DT6/End.DT46 SRv6 Endpoint Behavior. UPF2 decapsulates 377 the packet, performs a lookup in a specific table associated with 378 that mobile network and forwards the packet toward the data network 379 (DN). 381 5.1.2. Packet flow - Downlink 383 The downlink packet flow is as follows: 385 UPF2_in : (Z,A) 386 UPF2_out: (U2::, U1::2) (Z,A) -> H.Encaps.Red 387 UPF1_out: (U2::, gNB::1) (Z,A) -> End.MAP 388 gNB_out : (Z,A) -> End.DX4, End.DX6, End.DX2 390 When the packet arrives at the UPF2, the UPF2 maps that flow into a 391 PDU Session. This PDU Session is associated with the segment 392 endpoint . UPF2 performs a H.Encaps.Red operation, 393 encapsulating the packet into a new IPv6 header with no SRH since 394 there is only one SID. 396 Upon packet arrival on UPF1, the SID U1::2 is a local SID associated 397 with the End.MAP SRv6 Endpoint Behavior. It maps the SID to the next 398 anchoring point and replaces U1::2 by gNB::1, that belongs to the 399 next hop. 401 Upon packet arrival on gNB, the SID gNB::1 corresponds to an End.DX4, 402 End.DX6 or End.DX2 behavior (depending on the PDU Session Type). The 403 gNB decapsulates the packet, removing the IPv6 header and all its 404 extensions headers, and forwards the traffic toward the UE. 406 5.2. Enhanced Mode 408 Enhanced mode improves scalability, provides traffic engineering 409 capabilities, and allows service programming 410 [I-D.ietf-spring-sr-service-programming], thanks to the use of 411 multiple SIDs in the SID list (instead of a direct connectivity in 412 between UPFs with no intermediate waypoints as in Traditional Mode). 414 Thus, the main difference is that the SR policy MAY include SIDs for 415 traffic engineering and service programming in addition to the 416 anchoring SIDs at UPFs. 418 Additionally in this mode the operator may choose to aggregate 419 several devices under the same SID list (e.g., stationary residential 420 meters connected to the same cell) to improve scalability. 422 The gNB/UPF control-plane (N2/N4 interface) is unchanged, 423 specifically a single IPv6 address is provided to the gNB. A local 424 policy instructs the gNB to use SRv6. 426 The gNB MAY resolve the IP address received via the control plane 427 into a SID list using a mechanism like PCEP, DNS-lookup, LISP 428 control-plane or others. The resolution mechanism is out of the 429 scope of this document. 431 Note that the SIDs MAY use the arguments Args.Mob.Session if required 432 by the UPFs. 434 Figure 3 shows an Enhanced mode topology. The gNB and the UPF are 435 SR-aware. The Figure shows two service segments, S1 and C1. S1 436 represents a VNF in the network, and C1 represents an intermediate 437 router used for Traffic Engineering purposes to enforce a low-latency 438 path in the network. Note that neither S1 nor C1 are required to 439 have an N4 interface. 441 +----+ SRv6 _______ 442 SRv6 --| C1 |--[N3] / \ 443 +--+ +-----+ [N3] / +----+ \ +------+ [N6] / \ 444 |UE|----| gNB |-- SRv6 / SRv6 --| UPF1 |------\ DN / 445 +--+ +-----+ \ [N3]/ TE +------+ \_______/ 446 SRv6 node \ +----+ / SRv6 node 447 -| S1 |- 448 +----+ 449 SRv6 node 450 VNF 452 Figure 3: Enhanced mode - Example topology 454 5.2.1. Packet flow - Uplink 456 The uplink packet flow is as follows: 458 UE_out : (A,Z) 459 gNB_out : (gNB, S1)(U1::1, C1; SL=2)(A,Z)->H.Encaps.Red 460 S1_out : (gNB, C1)(U1::1, C1; SL=1)(A,Z) 461 C1_out : (gNB, U1::1)(A,Z) ->End with PSP 462 UPF1_out: (A,Z) ->End.DT4,End.DT6,End.DT2U 464 UE sends its packet (A,Z) on a specific bearer to its gNB. gNB's 465 control plane associates that session from the UE(A) with the IPv6 466 address B. gNB's control plane does a lookup on B to find the 467 related SID list . 469 When gNB transmits the packet, it contains all the segments of the SR 470 policy. The SR policy includes segments for traffic engineering (C1) 471 and for service programming (S1). 473 Nodes S1 and C1 perform their related Endpoint functionality and 474 forward the packet. 476 When the packet arrives at UPF1, the active segment (U1::1) is an 477 End.DT4/End.DT6/End.DT2U which performs the decapsulation (removing 478 the IPv6 header with all its extension headers) and forwards toward 479 the data network. 481 5.2.2. Packet flow - Downlink 483 The downlink packet flow is as follows: 485 UPF1_in : (Z,A) ->UPF1 maps the flow w/ 486 SID list 487 UPF1_out: (U1::1, C1)(gNB::1, S1; SL=2)(Z,A)->H.Encaps.Red 488 C1_out : (U1::1, S1)(gNB::1, S1; SL=1)(Z,A) 489 S1_out : (U1::1, gNB::1)(Z,A) ->End with PSP 490 gNB_out : (Z,A) ->End.DX4/End.DX6/End.DX2 492 When the packet arrives at the UPF1, the UPF1 maps that particular 493 flow into a UE PDU Session. This UE PDU Session is associated with 494 the policy . The UPF1 performs a H.Encaps.Red 495 operation, encapsulating the packet into a new IPv6 header with its 496 corresponding SRH. 498 The nodes C1 and S1 perform their related Endpoint processing. 500 Once the packet arrives at the gNB, the IPv6 DA corresponds to an 501 End.DX4, End.DX6 or End.DX2 behavior at the gNB (depending on the 502 underlying traffic). The gNB decapsulates the packet, removing the 503 IPv6 header, and forwards the traffic towards the UE. The SID gNB::1 504 is one example of a SID associated to this service. 506 Note that there are several means to provide the UE session 507 aggregation. The decision on which one to use is a local decision 508 made by the operator. One option is to use the Args.Mob.Session 509 (Section 6.1). Another option comprises the gNB performing an IP 510 lookup on the inner packet by using the End.DT4, End.DT6, and End.DT2 511 behaviors. 513 5.2.3. Scalability 515 The Enhanced Mode improves since it allows the aggregation of several 516 UEs under the same SID list. For example, in the case of stationary 517 residential meters that are connected to the same cell, all such 518 devices can share the same SID list. This improves scalability 519 compared to Traditional Mode (unique SID per UE) and compared to 520 GTP-U (dedicated TEID per UE). 522 5.3. Enhanced mode with unchanged gNB GTP behavior 524 This section describes two mechanisms for interworking with legacy 525 gNBs that still use GTP: one for IPv4, and another for IPv6. 527 In the interworking scenarios as illustrated in Figure 4, the gNB 528 does not support SRv6. The gNB supports GTP encapsulation over IPv4 529 or IPv6. To achieve interworking, an SR Gateway (SRGW) entity is 530 added. The SRGW maps the GTP traffic into SRv6. 532 The SRGW is not an anchor point and maintains very little state. For 533 this reason, both IPv4 and IPv6 methods scale to millions of UEs. 535 _______ 536 IP GTP SRv6 / \ 537 +--+ +-----+ [N3] +------+ [N9] +------+ [N6] / \ 538 |UE|------| gNB |------| SRGW |--------| UPF |---------\ DN / 539 +--+ +-----+ +------+ +------+ \_______/ 540 SR Gateway SRv6 node 542 Figure 4: Example topology for interworking 544 Both of the mechanisms described in this section are applicable to 545 either the Traditional Mode or the Enhanced Mode. 547 5.3.1. Interworking with IPv6 GTP 549 In this interworking mode the gNB at the N3 interface uses GTP over 550 IPv6. 552 Key points: 554 * The gNB is unchanged (control-plane or user-plane) and 555 encapsulates into GTP (N3 interface is not modified). 556 * The 5G Control-Plane towards the gNB (N2 interface) is unmodified; 557 one IPv6 address is needed per SLA(i.e. a BSID at the SRGW). The 558 SRv6 SID is different depending on the required SLA. 559 * In the uplink, the SRGW removes GTP, finds the SID list related to 560 the IPv6 DA, and adds SRH with the SID list. 561 * There is no state for the downlink at the SRGW. 562 * There is simple state in the uplink at the SRGW; using Enhanced 563 mode results in fewer SR policies on this node. An SR policy is 564 shared across UEs as long as they belong to the same context 565 (i.e., tenant). A set of many different policies (i.e., different 566 SLAs) increases the amount of state required. 567 * When a packet from the UE leaves the gNB, it is SR-routed. This 568 simplifies network slicing [I-D.ietf-lsr-flex-algo]. 570 * In the uplink, the SRv6 BSID steers traffic into an SR policy when 571 it arrives at the SRGW. 573 An example topology is shown in Figure 5. 575 S1 and C1 are two service segments. S1 represents a VNF in the 576 network, and C1 represents a router configured for Traffic 577 Engineering. 579 +----+ 580 IPv6/GTP -| S1 |- ___ 581 +--+ +-----+ [N3] / +----+ \ / 582 |UE|--| gNB |- SRv6 / SRv6 \ +----+ +------+ [N6] / 583 +--+ +-----+ \ [N9]/ VNF -| C1 |---| UPF2 |------\ DN 584 GTP \ +------+ / +----+ +------+ \___ 585 -| SRGW |- SRv6 SRv6 586 +------+ TE 587 SR Gateway 589 Figure 5: Enhanced mode with unchanged gNB IPv6/GTP behavior 591 5.3.1.1. Packet flow - Uplink 593 The uplink packet flow is as follows: 595 UE_out : (A,Z) 596 gNB_out : (gNB, B)(GTP: TEID T)(A,Z) -> Interface N3 unmodified 597 (IPv6/GTP) 598 SRGW_out: (SRGW, S1)(U2::T, C1; SL=2)(A,Z) -> B is an End.M.GTP6.D 599 SID at the SRGW 600 S1_out : (SRGW, C1)(U2::T, C1; SL=1)(A,Z) 601 C1_out : (SRGW, U2::T)(A,Z) -> End with PSP 602 UPF2_out: (A,Z) -> End.DT4 or End.DT6 604 The UE sends a packet destined to Z toward the gNB on a specific 605 bearer for that session. The gNB, which is unmodified, encapsulates 606 the packet into IPv6, UDP, and GTP headers. The IPv6 DA B, and the 607 GTP TEID T are the ones received in the N2 interface. 609 The IPv6 address that was signaled over the N2 interface for that UE 610 PDU Session, B, is now the IPv6 DA. B is an SRv6 Binding SID at the 611 SRGW. Hence the packet is routed to the SRGW. 613 When the packet arrives at the SRGW, the SRGW identifies B as an 614 End.M.GTP6.D Binding SID (see Section 6.3). Hence, the SRGW removes 615 the IPv6, UDP, and GTP headers, and pushes an IPv6 header with its 616 own SRH containing the SIDs bound to the SR policy associated with 617 this BindingSID. There at least one instance of the End.M.GTP6.D SID 618 per PDU type. 620 S1 and C1 perform their related Endpoint functionality and forward 621 the packet. 623 When the packet arrives at UPF2, the active segment is (U2::1) which 624 is bound to End.DT4/6. UPF2 then decapsulates (removing the outer 625 IPv6 header with all its extension headers) and forwards the packet 626 toward the data network. 628 5.3.1.2. Packet flow - Downlink 630 The downlink packet flow is as follows: 632 UPF2_in : (Z,A) -> UPF2 maps the flow with 633 634 UPF2_out: (U2::1, C1)(gNB, SRGW::TEID, S1; SL=3)(Z,A) -> H.Encaps.Red 635 C1_out : (U2::1, S1)(gNB, SRGW::TEID, S1; SL=2)(Z,A) 636 S1_out : (U2::1, SRGW::TEID)(gNB, SRGW::TEID, S1, SL=1)(Z,A) 637 SRGW_out: (SRGW, gNB)(GTP: TEID=T)(Z,A) -> SRGW/96 is End.M.GTP6.E 638 gNB_out : (Z,A) 640 When a packet destined to A arrives at the UPF2, the UPF2 performs a 641 lookup in the table associated to A and finds the SID list . The UPF2 performs an H.Encaps.Red operation, 643 encapsulating the packet into a new IPv6 header with its 644 corresponding SRH. 646 C1 and S1 perform their related Endpoint processing. 648 Once the packet arrives at the SRGW, the SRGW identifies the active 649 SID as an End.M.GTP6.E function. The SRGW removes the IPv6 header 650 and all its extensions headers. The SRGW generates new IPv6, UDP, 651 and GTP headers. The new IPv6 DA is the gNB which is the last SID in 652 the received SRH. The TEID in the generated GTP header is an 653 argument of the received End.M.GTP6.E SID. The SRGW pushes the 654 headers to the packet and forwards the packet toward the gNB. There 655 is one instance of the End.M.GTP6.E SID per PDU type. 657 Once the packet arrives at the gNB, the packet is a regular IPv6/GTP 658 packet. The gNB looks for the specific radio bearer for that TEID 659 and forward it on the bearer. This gNB behavior is not modified from 660 current and previous generations. 662 5.3.1.3. Scalability 664 For the downlink traffic, the SRGW is stateless. All the state is in 665 the SRH pushed by the UPF2. The UPF2 must have the UE states since 666 it is the UE's session anchor point. 668 For the uplink traffic, the state at the SRGW does not necessarily 669 need to be unique per PDU Session; the SR policy can be shared among 670 UEs. This enables more scalable SRGW deployments compared to a 671 solution holding millions of states, one or more per UE. 673 5.3.2. Interworking with IPv4 GTP 675 In this interworking mode the gNB uses GTP over IPv4 in the N3 676 interface 678 Key points: 680 * The gNB is unchanged and encapsulates packets into GTP (the N3 681 interface is not modified). 682 * In the uplink, traffic is classified by SRGW's classification 683 engine and steered into an SR policy. The SRGW may be implemented 684 in a UPF or as a separate entity. 685 * SRGW removes GTP, finds the SID list related to DA, and adds an 686 SRH with the SID list. 688 An example topology is shown in Figure 6. In this mode the gNB is an 689 unmodified gNB using IPv4/GTP. The UPFs are SR-aware. As before, 690 the SRGW maps the IPv4/GTP traffic to SRv6. 692 S1 and C1 are two service segment endpoints. S1 represents a VNF in 693 the network, and C1 represents a router configured for Traffic 694 Engineering. 696 +----+ 697 IPv4/GTP -| S1 |- ___ 698 +--+ +-----+ [N3] / +----+ \ / 699 |UE|--| gNB |- SRv6 / SRv6 \ +----+ +------+ [N6] / 700 +--+ +-----+ \ [N9]/ VNF -| C1 |---| UPF2 |------\ DN 701 GTP \ +------+ / +----+ +------+ \___ 702 -| UPF1 |- SRv6 SRv6 703 +------+ TE 704 SR Gateway 706 Figure 6: Enhanced mode with unchanged gNB IPv4/GTP behavior 708 5.3.2.1. Packet flow - Uplink 710 The uplink packet flow is as follows: 712 gNB_out : (gNB, B)(GTP: TEID T)(A,Z) -> Interface N3 713 unchanged IPv4/GTP 714 SRGW_out: (SRGW, S1)(U2::1, C1; SL=2)(A,Z) -> H.M.GTP4.D function 715 S1_out : (SRGW, C1)(U2::1, C1; SL=1)(A,Z) 716 C1_out : (SRGW, U2::1) (A,Z) -> PSP 717 UPF2_out: (A,Z) -> End.DT4 or End.DT6 719 The UE sends a packet destined to Z toward the gNB on a specific 720 bearer for that session. The gNB, which is unmodified, encapsulates 721 the packet into a new IPv4, UDP, and GTP headers. The IPv4 DA, B, 722 and the GTP TEID are the ones received at the N2 interface. 724 When the packet arrives at the SRGW for UPF1, the SRGW has an 725 classification engine rule for incoming traffic from the gNB, that 726 steers the traffic into an SR policy by using the function 727 H.M.GTP4.D. The SRGW removes the IPv4, UDP, and GTP headers and 728 pushes an IPv6 header with its own SRH containing the SIDs related to 729 the SR policy associated with this traffic. The SRGW forwards 730 according to the new IPv6 DA. 732 S1 and C1 perform their related Endpoint functionality and forward 733 the packet. 735 When the packet arrives at UPF2, the active segment is (U2::1) which 736 is bound to End.DT4/6 which performs the decapsulation (removing the 737 outer IPv6 header with all its extension headers) and forwards toward 738 the data network. 740 Note that the interworking mechanisms for IPv4/GTP and IPv6/GTP 741 differs. This is due to the fact that in IPv6/GTP we can leverage 742 the remote steering capabilities provided by the Segment Routing 743 BSID. In IPv4 this construct is not available, and building a 744 similar mechanism would require a significant address consumption. 746 5.3.2.2. Packet flow - Downlink 748 The downlink packet flow is as follows: 750 UPF2_in : (Z,A) -> UPF2 maps flow with SID 751 752 UPF2_out: (U2::1, C1)(GW::SA:DA:TEID, S1; SL=2)(Z,A) ->H.Encaps.Red 753 C1_out : (U2::1, S1)(GW::SA:DA:TEID, S1; SL=1)(Z,A) 754 S1_out : (U2::1, GW::SA:DA:TEID)(Z,A) 755 SRGW_out: (GW, gNB)(GTP: TEID=T)(Z,A) -> End.M.GTP4.E 756 gNB_out : (Z,A) 758 When a packet destined to A arrives at the UPF2, the UPF2 performs a 759 lookup in the table associated to A and finds the SID list . The UPF2 performs a H.Encaps.Red operation, 761 encapsulating the packet into a new IPv6 header with its 762 corresponding SRH. 764 The nodes C1 and S1 perform their related Endpoint processing. 766 Once the packet arrives at the SRGW, the SRGW identifies the active 767 SID as an End.M.GTP4.E function. The SRGW removes the IPv6 header 768 and all its extensions headers. The SRGW generates an IPv4, UDP, and 769 GTP headers. The IPv4 SA and DA are received as SID arguments. The 770 TEID in the generated GTP header is also the arguments of the 771 received End.M.GTP4.E SID. The SRGW pushes the headers to the packet 772 and forwards the packet toward the gNB. 774 When the packet arrives at the gNB, the packet is a regular IPv4/GTP 775 packet. The gNB looks for the specific radio bearer for that TEID 776 and forwards it on the bearer. This gNB behavior is not modified 777 from current and previous generations. 779 5.3.2.3. Scalability 781 For the downlink traffic, the SRGW is stateless. All the state is in 782 the SRH pushed by the UPF2. The UPF must have this UE-base state 783 anyway (since it is its anchor point). 785 For the uplink traffic, the state at the SRGW is dedicated on a per 786 UE/session basis according to a classification engine. There is 787 state for steering the different sessions in the form of an SR 788 Policy. However, SR policies are shared among several UE/sessions. 790 5.3.3. Extensions to the interworking mechanisms 792 In this section we presented two mechanisms for interworking with 793 gNBs and UPFs that do not support SRv6. These mechanisms are used to 794 support GTP over IPv4 and IPv6. 796 Even though we have presented these methods as an extension to the 797 "Enhanced mode", it is straightforward in its applicability to the 798 "Traditional mode". 800 5.4. SRv6 Drop-in Interworking 802 In this section we introduce another mode useful for legacy gNB and 803 UPFs that still operate with GTP-U. This mode provides an 804 SRv6-enabled user plane in between two GTP-U tunnel endpoints. 806 In this mode we employ two SRGWs that map GTP-U traffic to SRv6 and 807 vice-versa. 809 Unlike other interworking modes, in this mode both of the mobility 810 overlay endpoints use GTP-U. Two SRGWs are deployed in either N3 or 811 N9 interface to realize an intermediate SR policy. 813 +----+ 814 -| S1 |- 815 +-----+ / +----+ \ 816 | gNB |- SRv6 / SRv6 \ +----+ +--------+ +-----+ 817 +-----+ \ / VNF -| C1 |---| SRGW-B |----| UPF | 818 GTP[N3]\ +--------+ / +----+ +--------+ +-----+ 819 -| SRGW-A |- SRv6 SR Gateway-B GTP 820 +--------+ TE 821 SR Gateway-A 823 Figure 7: Example topology for SRv6 Drop-in mode 825 The packet flow of Figure 7 is as follows: 827 gNB_out : (gNB, U::1)(GTP: TEID T)(A,Z) 828 GW-A_out: (GW-A, S1)(U::1, SGB::TEID, C1; SL=3)(A,Z)->U::1 is an 829 End.M.GTP6.D.Di 830 SID at SRGW-A 831 S1_out : (GW-A, C1)(U::1, SGB::TEID, C1; SL=2)(A,Z) 832 C1_out : (GW-A, SGB::TEID)(U::1, SGB::TEID, C1; SL=1)(A,Z) 833 GW-B_out: (GW-B, U::1)(GTP: TEID T)(A,Z) ->SGB::TEID is an 834 End.M.GTP6.E 835 SID at SRGW-B 836 UPF_out : (A,Z) 838 When a packet destined to Z is sent to the gNB, which is unmodified 839 (control-plane and user-plane remain GTP-U), gNB performs 840 encapsulation into a new IP, UDP, and GTP headers. The IPv6 DA, 841 U::1, and the GTP TEID are the ones received at the N2 interface. 843 The IPv6 address that was signaled over the N2 interface for that PDU 844 Session, U::1, is now the IPv6 DA. U::1 is an SRv6 Binding SID at 845 SRGW-A. Hence the packet is routed to the SRGW. 847 When the packet arrives at SRGW-A, the SRGW identifies U::1 as an 848 End.M.GTP6.D.Di Binding SID (see Section 6.4). Hence, the SRGW 849 removes the IPv6, UDP, and GTP headers, and pushes an IPv6 header 850 with its own SRH containing the SIDs bound to the SR policy 851 associated with this Binding SID. There is one instance of the 852 End.M.GTP6.D.Di SID per PDU type. 854 S1 and C1 perform their related Endpoint functionality and forward 855 the packet. 857 Once the packet arrives at SRGW-B, the SRGW identifies the active SID 858 as an End.M.GTP6.E function. The SRGW removes the IPv6 header and 859 all its extensions headers. The SRGW generates new IPv6, UDP, and 860 GTP headers. The new IPv6 DA is U::1 which is the last SID in the 861 received SRH. The TEID in the generated GTP header is an argument of 862 the received End.M.GTP6.E SID. The SRGW pushes the headers to the 863 packet and forwards the packet toward UPF. There is one instance of 864 the End.M.GTP6.E SID per PDU type. 866 Once the packet arrives at UPF, the packet is a regular IPv6/GTP 867 packet. The UPF looks for the specific rule for that TEID to forward 868 the packet. This UPF behavior is not modified from current and 869 previous generations. 871 6. SRv6 Segment Endpoint Mobility Behaviors 873 6.1. Args.Mob.Session 875 Args.Mob.Session provide per-session information for charging, 876 buffering and lawful intercept (among others) required by some mobile 877 nodes. The Args.Mob.Session argument format is used in combination 878 with End.Map, End.DT4/End.DT6/End.DT46 and End.DX4/End.DX6/End.DX2 879 behaviors. Note that proposed format is applicable for 5G networks, 880 while similar formats could be used for legacy networks. 882 0 1 2 3 883 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 1 884 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 885 | QFI |R|U| PDU Session ID | 886 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 887 |PDU Sess(cont')| 888 +-+-+-+-+-+-+-+-+ 890 Figure 8: Args.Mob.Session format 892 * QFI: QoS Flow Identifier [TS.38415] 893 * R: Reflective QoS Indication [TS.23501]. This parameter indicates 894 the activation of reflective QoS towards the UE for the 895 transferred packet. Reflective QoS enables the UE to map UL User 896 Plane traffic to QoS Flows without SMF provided QoS rules. 897 * U: Unused and for future use. MUST be 0 on transmission and 898 ignored on receipt. 899 * PDU Session ID: Identifier of PDU Session. The GTP-U equivalent 900 is TEID. 902 Arg.Mob.Session is required in case that one SID aggregates multiple 903 PDU Sessions. Since the SRv6 SID is likely NOT to be instantiated 904 per PDU session, Args.Mob.Session helps the UPF to perform the 905 behaviors which require per QFI and/or per PDU Session granularity. 907 Note that the encoding of user-plane messages (e.g., Echo Request, 908 Echo Reply, Error Indication and End Marker) is out of the scope of 909 this draft. [I-D.murakami-dmm-user-plane-message-encoding] defines 910 one possible encoding. 912 6.2. End.MAP 914 The "Endpoint behavior with SID mapping" behavior (End.MAP for short) 915 is used in several scenarios. Particularly in mobility, End.MAP is 916 used by the intermediate UPFs. 918 When node N receives a packet whose IPv6 DA is S and S is a local 919 End.MAP SID, N does: 921 S01. If (IPv6 Hop Limit <= 1) { 922 S02. Send an ICMP Time Exceeded message to the Source Address, 923 Code 0 (Hop limit exceeded in transit), 924 interrupt packet processing, and discard the packet. 925 S03. } 926 S04. Decrement IPv6 Hop Limit by 1 927 S05. Update the IPv6 DA with the new mapped SID 928 S06. Submit the packet to the egress IPv6 FIB lookup for 929 transmission to the new destination 931 Notes: The SIDs in the SRH are not modified. 933 6.3. End.M.GTP6.D 935 The "Endpoint behavior with IPv6/GTP decapsulation into SR policy" 936 behavior (End.M.GTP6.D for short) is used in interworking scenario 937 for the uplink towards SRGW from the legacy gNB using IPv6/GTP. Any 938 SID instance of this behavior is associated with an SR Policy B and 939 an IPv6 Source Address A. 941 When the SR Gateway node N receives a packet destined to S and S is a 942 local End.M.GTP6.D SID, N does: 944 S01. When an SRH is processed { 945 S02. If (Segments Left != 0) { 946 S03. Send an ICMP Parameter Problem to the Source Address, 947 Code 0 (Erroneous header field encountered), 948 Pointer set to the Segments Left field, 949 interrupt packet processing, and discard the packet. 950 S04. } 951 S05. Proceed to process the next header in the packet 952 S06. } 954 When processing the Upper-layer header of a packet matching a FIB 955 entry locally instantiated as an End.M.GTP6.D SID, N does: 957 S01. If (Next Header (NH) == UDP & UDP_Dest_port == GTP) { 958 S02. Copy the GTP TEID and QFI to buffer memory 959 S03. Pop the IPv6, UDP, and GTP Headers 960 S04. Push a new IPv6 header with its own SRH containing B 961 S05. Set the outer IPv6 SA to A 962 S06. Set the outer IPv6 DA to the first SID of B 963 S07. Set the outer Payload Length, Traffic Class, Flow Label, 964 Hop Limit, and Next-Header (NH) fields 965 S08. Write in the last SID of the SRH the Args.Mob.Session 966 based on the information of buffer memory 967 S09. Submit the packet to the egress IPv6 FIB lookup and 968 transmission to the new destination 969 S10. } Else { 970 S11. Process as per [RFC8986] Section 4.1.1 971 S12. } 973 Notes: The NH is set based on the SID parameter. There is one 974 instantiation of the End.M.GTP6.D SID per PDU Session Type, hence the 975 NH is already known in advance. For the IPv4v6 PDU Session Type, in 976 addition we inspect the first nibble of the PDU to know the NH value. 978 The prefix of last segment (S3 in above example) SHOULD be followed 979 by an Arg.Mob.Session argument space which is used to provide the 980 session identifiers. 982 6.4. End.M.GTP6.D.Di 984 The "Endpoint behavior with IPv6/GTP decapsulation into SR policy for 985 Drop-in Mode" behavior (End.M.GTP6.D.Di for short) is used in SRv6 986 drop-in interworking scenario described in Section 5.4. The 987 difference between End.M.GTP6.D as another variant of IPv6/GTP 988 decapsulation function is that the original IPv6 DA of GTP packet is 989 preserved as the last SID in SRH. 991 Any SID instance of this behavior is associated with an SR Policy B 992 and an IPv6 Source Address A. 994 When the SR Gateway node N receives a packet destined to S and S is a 995 local End.M.GTP6.D.Di SID, N does: 997 S01. When an SRH is processed { 998 S02. If (Segments Left != 0) { 999 S03. Send an ICMP Parameter Problem to the Source Address, 1000 Code 0 (Erroneous header field encountered), 1001 Pointer set to the Segments Left field, 1002 interrupt packet processing, and discard the packet. 1003 S04. } 1004 S05. Proceed to process the next header in the packet 1005 S06. } 1007 When processing the Upper-layer header of a packet matching a FIB 1008 entry locally instantiated as an End.M.GTP6.Di SID, N does: 1010 S01. If (Next Header = UDP & UDP_Dest_port = GTP) { 1011 S02. Copy S to buffer memory 1012 S03. Pop the IPv6, UDP, and GTP Headers 1013 S04. Push a new IPv6 header with its own SRH containing B 1014 S05. Set the outer IPv6 SA to A 1015 S06. Set the outer IPv6 DA to the first SID of B 1016 S07. Set the outer Payload Length, Traffic Class, Flow Label, 1017 Hop Limit, and Next-Header fields 1018 S08. Write S to the SRH 1019 S09. Submit the packet to the egress IPv6 FIB lookup and 1020 transmission to the new destination 1021 S10. } Else { 1022 S11. Process as per [RFC8986] Section 4.1.1 1023 S12. } 1025 Notes: The NH is set based on the SID parameter. There is one 1026 instantiation of the End.M.GTP6.D SID per PDU Session Type, hence the 1027 NH is already known in advance. For the IPv4v6 PDU Session Type, in 1028 addition we inspect the first nibble of the PDU to know the NH value. 1030 The prefix of A SHOULD be an End.M.GTP6.E SID instantiated at an SR 1031 gateway. 1033 6.5. End.M.GTP6.E 1035 The "Endpoint behavior with encapsulation for IPv6/GTP tunnel" 1036 behavior (End.M.GTP6.E for short) is used among others in the 1037 interworking scenario for the downlink toward the legacy gNB using 1038 IPv6/GTP. 1040 The prefix of End.M.GTP6.E SID MUST be followed by the 1041 Arg.Mob.Session argument space which is used to provide the session 1042 identifiers. 1044 When the SR Gateway node N receives a packet destined to S, and S is 1045 a local End.M.GTP6.E SID, N does the following: 1047 S01. When an SRH is processed { 1048 S02. If (Segments Left != 1) { 1049 S03. Send an ICMP Parameter Problem to the Source Address, 1050 Code 0 (Erroneous header field encountered), 1051 Pointer set to the Segments Left field, 1052 interrupt packet processing, and discard the packet. 1053 S04. } 1054 S05. Proceed to process the next header in the packet 1055 S06. } 1057 When processing the Upper-layer header of a packet matching a FIB 1058 entry locally instantiated as an End.M.GTP6.E SID, N does: 1060 S01. Copy SRH[0] and S to buffer memory 1061 S02. Pop the IPv6 header and all its extension headers 1062 S03. Push a new IPv6 header with a UDP/GTP Header 1063 S04. Set the outer IPv6 SA to A 1064 S05. Set the outer IPv6 DA from buffer memory 1065 S06. Set the outer Payload Length, Traffic Class, Flow Label, 1066 Hop Limit, and Next-Header fields 1067 S07. Set the GTP TEID (from buffer memory) 1068 S08. Submit the packet to the egress IPv6 FIB lookup and 1069 transmission to the new destination 1070 S09. } 1072 Notes: An End.M.GTP6.E SID MUST always be the penultimate SID. The 1073 TEID is extracted from the argument space of the current SID. 1075 The source address A SHOULD be an End.M.GTP6.D SID instantiated at an 1076 SR gateway. 1078 6.6. End.M.GTP4.E 1080 The "Endpoint behavior with encapsulation for IPv4/GTP tunnel" 1081 behavior (End.M.GTP4.E for short) is used in the downlink when doing 1082 interworking with legacy gNB using IPv4/GTP. 1084 When the SR Gateway node N receives a packet destined to S and S is a 1085 local End.M.GTP4.E SID, N does: 1087 S01. When an SRH is processed { 1088 S02. If (Segments Left != 0) { 1089 S03. Send an ICMP Parameter Problem to the Source Address, 1090 Code 0 (Erroneous header field encountered), 1091 Pointer set to the Segments Left field, 1092 interrupt packet processing, and discard the packet. 1093 S04. } 1094 S05. Proceed to process the next header in the packet 1095 S06. } 1097 When processing the Upper-layer header of a packet matching a FIB 1098 entry locally instantiated as an End.M.GTP4.E SID, N does: 1100 S01. If (Next Header = UDP & UDP_Dest_port = GTP) { 1101 S02. Store the IPv6 DA and SA in buffer memory 1102 S03. Pop the IPv6 header and all its extension headers 1103 S04. Push a new IPv4 header with a UDP/GTP Header 1104 S05. Set the outer IPv4 SA and DA (from buffer memory) 1105 S06. Set the outer Total Length, DSCP, Time To Live, and 1106 Next-Header fields 1107 S07. Set the GTP TEID (from buffer memory) 1108 S08. Submit the packet to the egress IPv6 FIB lookup and 1109 transmission to the new destination 1110 S09. } Else { 1111 S10. Process as per [NET-PGM] Section 4.1.1 1112 S11. } 1114 Notes: The End.M.GTP4.E SID in S has the following format: 1116 0 127 1117 +-----------------------+-------+----------------+---------+ 1118 | SRGW-IPv6-LOC-FUNC |IPv4DA |Args.Mob.Session|0 Padded | 1119 +-----------------------+-------+----------------+---------+ 1120 128-a-b-c a b c 1122 Figure 9: End.M.GTP4.E SID Encoding 1124 The IPv6 Source Address has the following format: 1126 0 127 1127 +----------------------+--------+--------------------------+ 1128 | Source UPF Prefix |IPv4 SA | any bit pattern(ignored) | 1129 +----------------------+--------+--------------------------+ 1130 128-a-b a b 1132 Figure 10: IPv6 SA Encoding for End.M.GTP4.E 1134 6.7. H.M.GTP4.D 1136 The "SR Policy Headend with tunnel decapsulation and map to an SRv6 1137 policy" behavior (H.M.GTP4.D for short) is used in the direction from 1138 legacy IPv4 user-plane to SRv6 user-plane network. 1140 When the SR Gateway node N receives a packet destined to a IW- 1141 IPv4-Prefix, N does: 1143 S01. IF Payload == UDP/GTP THEN 1144 S02. Pop the outer IPv4 header and UDP/GTP headers 1145 S03. Copy IPv4 DA, TEID to form SID B 1146 S04. Copy IPv4 SA to form IPv6 SA B' 1147 S05. Encapsulate the packet into a new IPv6 header ;;Ref1 1148 S06. Set the IPv6 DA = B 1149 S07. Forward along the shortest path to B 1150 S08. ELSE 1151 S09. Drop the packet 1153 Ref1: The NH value is identified by inspecting the first nibble of 1154 the inner payload. 1156 The SID B has the following format: 1158 0 127 1159 +-----------------------+-------+----------------+---------+ 1160 |Destination UPF Prefix |IPv4DA |Args.Mob.Session|0 Padded | 1161 +-----------------------+-------+----------------+---------+ 1162 128-a-b-c a b c 1164 Figure 11: H.M.GTP4.D SID Encoding 1166 The SID B MAY be an SRv6 Binding SID instantiated at the first UPF 1167 (U1) to bind an SR policy [I-D.ietf-spring-segment-routing-policy]. 1169 6.8. End.Limit: Rate Limiting behavior 1171 The mobile user-plane requires a rate-limit feature. For this 1172 purpose, we define a new behavior "End.Limit". The "End.Limit" 1173 behavior encodes in its arguments the rate limiting parameter that 1174 should be applied to this packet. Multiple flows of packets should 1175 have the same group identifier in the SID when those flows are in the 1176 same AMBR (Aggregate Maximum Bit Rate) group. The encoding format of 1177 the rate limit segment SID is as follows: 1179 +----------------------+----------+-----------+ 1180 | LOC+FUNC rate-limit | group-id | limit-rate| 1181 +----------------------+----------+-----------+ 1182 128-i-j i j 1184 Figure 12: End.Limit: Rate limiting behavior argument format 1186 If the limit-rate bits are set to zero, the node should not do rate 1187 limiting unless static configuration or control-plane sets the limit 1188 rate associated to the SID. 1190 7. SRv6 supported 3GPP PDU session types 1192 The 3GPP [TS.23501] defines the following PDU session types: 1194 * IPv4 1195 * IPv6 1196 * IPv4v6 1197 * Ethernet 1198 * Unstructured 1200 SRv6 supports the 3GPP PDU session types without any protocol 1201 overhead by using the corresponding SRv6 behaviors (End.DX4, End.DT4 1202 for IPv4 PDU sessions; End.DX6, End.DT6, End.T for IPv6 PDU sessions; 1203 End.DT46 for IPv4v6 PDU sessions; End.DX2 for L2 and Unstructured PDU 1204 sessions). 1206 8. Network Slicing Considerations 1208 A mobile network may be required to implement "network slices", which 1209 logically separate network resources. User-plane behaviors 1210 represented as SRv6 segments would be part of a slice. 1212 [I-D.ietf-spring-segment-routing-policy] describes a solution to 1213 build basic network slices with SR. Depending on the requirements, 1214 these slices can be further refined by adopting the mechanisms from: 1216 * IGP Flex-Algo [I-D.ietf-lsr-flex-algo] 1217 * Inter-Domain policies 1218 [I-D.ietf-spring-segment-routing-central-epe] 1220 Furthermore, these can be combined with ODN/AS (On Demand Nexthop/ 1221 Automated Steering) [I-D.ietf-spring-segment-routing-policy] for 1222 automated slice provisioning and traffic steering. 1224 Further details on how these tools can be used to create end to end 1225 network slices are documented in 1226 [I-D.ali-spring-network-slicing-building-blocks]. 1228 9. Control Plane Considerations 1230 This document focuses on user-plane behavior and its independence 1231 from the control plane. While there are benefits in an enhanced 1232 control plane (e.g., to dynamically configure SR policies from a 1233 controller, session aggregation), this document does not impose any 1234 change to the existent mobility control plane. 1236 Section 11 allocates SRv6 Segment Endpoint Behavior codepoints for 1237 the new behaviors defined in this document. 1239 10. Security Considerations 1241 The security considerations for Segment Routing are discussed in 1242 [RFC8402]. More specifically for SRv6 the security considerations 1243 and the mechanisms for securing an SR domain are discussed in 1244 [RFC8754]. Together, they describe the required security mechanisms 1245 that allow establishment of an SR domain of trust to operate 1246 SRv6-based services for internal traffic while preventing any 1247 external traffic from accessing or exploiting the SRv6-based 1248 services. 1250 The technology described in this document is applied to a mobile 1251 network that is within the SR Domain. 1253 This document introduces new SRv6 Endpoint Behaviors. Those 1254 behaviors do not need any special security consideration given that 1255 it is deployed within that SR Domain. 1257 11. IANA Considerations 1259 The following values have been allocated within the "SRv6 Endpoint 1260 Behaviors" [RFC8986] sub-registry belonging to the top-level "Segment 1261 Routing Parameters" registry: 1263 +=======+========+===================+===========+ 1264 | Value | Hex | Endpoint behavior | Reference | 1265 +=======+========+===================+===========+ 1266 | 40 | 0x0028 | End.MAP | [This.ID] | 1267 +-------+--------+-------------------+-----------+ 1268 | 41 | 0x0029 | End.Limit | [This.ID] | 1269 +-------+--------+-------------------+-----------+ 1270 | 69 | 0x0045 | End.M.GTP6.D | [This.ID] | 1271 +-------+--------+-------------------+-----------+ 1272 | 70 | 0x0046 | End.M.GTP6.Di | [This.ID] | 1273 +-------+--------+-------------------+-----------+ 1274 | 71 | 0x0047 | End.M.GTP6.E | [This.ID] | 1275 +-------+--------+-------------------+-----------+ 1276 | 72 | 0x0048 | End.M.GTP4.E | [This.ID] | 1277 +-------+--------+-------------------+-----------+ 1279 Table 1: SRv6 Mobile User-plane Endpoint 1280 Behavior Types 1282 12. Acknowledgements 1284 The authors would like to thank Daisuke Yokota, Bart Peirens, 1285 Ryokichi Onishi, Kentaro Ebisawa, Peter Bosch, Darren Dukes, Francois 1286 Clad, Sri Gundavelli, Sridhar Bhaskaran, Arashmid Akhavain, Ravi 1287 Shekhar, Aeneas Dodd-Noble, Carlos Jesus Bernardos, Dirk v. Hugo and 1288 Jeffrey Zhang for their useful comments of this work. 1290 13. Contributors 1292 Kentaro Ebisawa Toyota Motor Corporation Japan 1294 Email: ebisawa@toyota-tokyo.tech 1296 Tetsuya Murakami Arrcus, Inc. United States of America 1298 Email: tetsuya.ietf@gmail.com 1300 14. References 1302 14.1. Normative References 1304 [I-D.ietf-spring-segment-routing-policy] 1305 Filsfils, C., Talaulikar, K., Voyer, D., Bogdanov, A., and 1306 P. Mattes, "Segment Routing Policy Architecture", Work in 1307 Progress, Internet-Draft, draft-ietf-spring-segment- 1308 routing-policy-13, 28 May 2021, 1309 . 1312 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 1313 Requirement Levels", BCP 14, RFC 2119, 1314 DOI 10.17487/RFC2119, March 1997, 1315 . 1317 [RFC8402] Filsfils, C., Ed., Previdi, S., Ed., Ginsberg, L., 1318 Decraene, B., Litkowski, S., and R. Shakir, "Segment 1319 Routing Architecture", RFC 8402, DOI 10.17487/RFC8402, 1320 July 2018, . 1322 [RFC8754] Filsfils, C., Ed., Dukes, D., Ed., Previdi, S., Leddy, J., 1323 Matsushima, S., and D. Voyer, "IPv6 Segment Routing Header 1324 (SRH)", RFC 8754, DOI 10.17487/RFC8754, March 2020, 1325 . 1327 [RFC8986] Filsfils, C., Ed., Camarillo, P., Ed., Leddy, J., Voyer, 1328 D., Matsushima, S., and Z. Li, "Segment Routing over IPv6 1329 (SRv6) Network Programming", RFC 8986, 1330 DOI 10.17487/RFC8986, February 2021, 1331 . 1333 [TS.23501] 3GPP, "System Architecture for the 5G System", 3GPP TS 1334 23.501 15.0.0, November 2017. 1336 14.2. Informative References 1338 [I-D.ali-spring-network-slicing-building-blocks] 1339 Ali, Z., Filsfils, C., Camarillo, P., and D. Voyer, 1340 "Building blocks for Slicing in Segment Routing Network", 1341 Work in Progress, Internet-Draft, draft-ali-spring- 1342 network-slicing-building-blocks-04, 21 February 2021, 1343 . 1346 [I-D.camarilloelmalky-springdmm-srv6-mob-usecases] 1347 Garvia, P. C., Filsfils, C., Elmalky, H., Matsushima, S., 1348 Voyer, D., Cui, A., and B. Peirens, "SRv6 Mobility Use- 1349 Cases", Work in Progress, Internet-Draft, draft- 1350 camarilloelmalky-springdmm-srv6-mob-usecases-02, 15 August 1351 2019, . 1354 [I-D.ietf-lsr-flex-algo] 1355 Psenak, P., Hegde, S., Filsfils, C., Talaulikar, K., and 1356 A. Gulko, "IGP Flexible Algorithm", Work in Progress, 1357 Internet-Draft, draft-ietf-lsr-flex-algo-17, 6 July 2021, 1358 . 1361 [I-D.ietf-spring-segment-routing-central-epe] 1362 Filsfils, C., Previdi, S., Dawra, G., Aries, E., and D. 1363 Afanasiev, "Segment Routing Centralized BGP Egress Peer 1364 Engineering", Work in Progress, Internet-Draft, draft- 1365 ietf-spring-segment-routing-central-epe-10, 21 December 1366 2017, . 1369 [I-D.ietf-spring-sr-service-programming] 1370 Clad, F., Xu, X., Filsfils, C., Bernier, D., Li, C., 1371 Decraene, B., Ma, S., Yadlapalli, C., Henderickx, W., and 1372 S. Salsano, "Service Programming with Segment Routing", 1373 Work in Progress, Internet-Draft, draft-ietf-spring-sr- 1374 service-programming-05, 10 September 2021, 1375 . 1378 [I-D.kohno-dmm-srv6mob-arch] 1379 Kohno, M., Clad, F., Camarillo, P., and Z. Ali, 1380 "Architecture Discussion on SRv6 Mobile User plane", Work 1381 in Progress, Internet-Draft, draft-kohno-dmm-srv6mob-arch- 1382 04, 6 May 2021, . 1385 [I-D.matsushima-spring-srv6-deployment-status] 1386 Matsushima, S., Filsfils, C., Ali, Z., Li, Z., and K. 1387 Rajaraman, "SRv6 Implementation and Deployment Status", 1388 Work in Progress, Internet-Draft, draft-matsushima-spring- 1389 srv6-deployment-status-11, 17 February 2021, 1390 . 1393 [I-D.murakami-dmm-user-plane-message-encoding] 1394 Murakami, T., Matsushima, S., Ebisawa, K., Camarillo, P., 1395 and R. Shekhar, "User Plane Message Encoding", Work in 1396 Progress, Internet-Draft, draft-murakami-dmm-user-plane- 1397 message-encoding-04, 2 September 2021, 1398 . 1401 [TS.29281] 3GPP, "General Packet Radio System (GPRS) Tunnelling 1402 Protocol User Plane (GTPv1-U)", 3GPP TS 29.281 15.1.0, 1403 December 2017. 1405 [TS.38415] 3GPP, "Draft Specification for 5GS container (TS 38.415)", 1406 3GPP R3-174510 0.0.0, August 2017. 1408 Appendix A. Implementations 1410 This document introduces new SRv6 Endpoint Behaviors. These 1411 behaviors have an open-source P4 implementation available in 1412 https://github.com/ebiken/p4srv6. 1414 Additionally, a full implementation of this document is available in 1415 Linux Foundation FD.io VPP project since release 20.05. More 1416 information available here: https://docs.fd.io/vpp/20.05/d7/d3c/ 1417 srv6_mobile_plugin_doc.html. 1419 There are also experimental implementations in M-CORD NGIC and Open 1420 Air Interface (OAI). 1422 Authors' Addresses 1424 Satoru Matsushima (editor) 1425 SoftBank 1426 Japan 1428 Email: satoru.matsushima@g.softbank.co.jp 1430 Clarence Filsfils 1431 Cisco Systems, Inc. 1432 Belgium 1434 Email: cf@cisco.com 1436 Miya Kohno 1437 Cisco Systems, Inc. 1438 Japan 1440 Email: mkohno@cisco.com 1442 Pablo Camarillo Garvia (editor) 1443 Cisco Systems, Inc. 1444 Spain 1446 Email: pcamaril@cisco.com 1448 Daniel Voyer 1449 Bell Canada 1450 Canada 1451 Email: daniel.voyer@bell.ca 1453 Charles E. Perkins 1454 Lupin Lodge 1455 20600 Aldercroft Heights Rd. 1456 Los Gatos, CA 95033 1457 United States of America 1459 Email: charliep@computer.org