idnits 2.17.1 draft-ietf-softwire-lw4over6-13.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- == There are 1 instance of lines with non-RFC6890-compliant IPv4 addresses in the document. If these are example addresses, they should be changed. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (November 13, 2014) is 3452 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(See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) == Outdated reference: A later version (-12) exists of draft-ietf-softwire-map-dhcp-10 == Outdated reference: A later version (-09) exists of draft-ietf-dhc-dynamic-shared-v4allocation-02 == Outdated reference: A later version (-13) exists of draft-ietf-pcp-port-set-07 == Outdated reference: A later version (-13) exists of draft-ietf-softwire-map-11 == Outdated reference: A later version (-08) exists of draft-ietf-softwire-unified-cpe-01 Summary: 0 errors (**), 0 flaws (~~), 7 warnings (==), 1 comment (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Softwire Working Group Y. Cui 3 Internet-Draft Tsinghua University 4 Intended status: Standards Track Q. Sun 5 Expires: May 17, 2015 China Telecom 6 M. Boucadair 7 France Telecom 8 T. Tsou 9 Huawei Technologies 10 Y. Lee 11 Comcast 12 I. Farrer 13 Deutsche Telekom AG 14 November 13, 2014 16 Lightweight 4over6: An Extension to the DS-Lite Architecture 17 draft-ietf-softwire-lw4over6-13 19 Abstract 21 Dual-Stack Lite (RFC 6333) describes an architecture for transporting 22 IPv4 packets over an IPv6 network. This document specifies an 23 extension to DS-Lite called Lightweight 4over6 which moves the 24 Network Address and Port Translation (NAPT) function from the 25 centralized DS-Lite tunnel concentrator to the tunnel client located 26 in the Customer Premises Equipment (CPE). This removes the 27 requirement for a Carrier Grade NAT function in the tunnel 28 concentrator and reduces the amount of centralized state that must be 29 held to a per-subscriber level. In order to delegate the NAPT 30 function and make IPv4 Address sharing possible, port-restricted IPv4 31 addresses are allocated to the CPEs. 33 Status of This Memo 35 This Internet-Draft is submitted in full conformance with the 36 provisions of BCP 78 and BCP 79. 38 Internet-Drafts are working documents of the Internet Engineering 39 Task Force (IETF). Note that other groups may also distribute 40 working documents as Internet-Drafts. The list of current Internet- 41 Drafts is at http://datatracker.ietf.org/drafts/current/. 43 Internet-Drafts are draft documents valid for a maximum of six months 44 and may be updated, replaced, or obsoleted by other documents at any 45 time. It is inappropriate to use Internet-Drafts as reference 46 material or to cite them other than as "work in progress." 48 This Internet-Draft will expire on May 17, 2015. 50 Copyright Notice 52 Copyright (c) 2014 IETF Trust and the persons identified as the 53 document authors. All rights reserved. 55 This document is subject to BCP 78 and the IETF Trust's Legal 56 Provisions Relating to IETF Documents 57 (http://trustee.ietf.org/license-info) in effect on the date of 58 publication of this document. Please review these documents 59 carefully, as they describe your rights and restrictions with respect 60 to this document. Code Components extracted from this document must 61 include Simplified BSD License text as described in Section 4.e of 62 the Trust Legal Provisions and are provided without warranty as 63 described in the Simplified BSD License. 65 Table of Contents 67 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 2 68 2. Conventions . . . . . . . . . . . . . . . . . . . . . . . . . 4 69 3. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 4 70 4. Lightweight 4over6 Architecture . . . . . . . . . . . . . . . 5 71 5. Lightweight B4 Behavior . . . . . . . . . . . . . . . . . . . 7 72 5.1. Lightweight B4 Provisioning with DHCPv6 . . . . . . . . . 7 73 5.2. Lightweight B4 Data Plane Behavior . . . . . . . . . . . 9 74 5.2.1. Fragmentation Behaviour . . . . . . . . . . . . . . . 11 75 6. Lightweight AFTR Behavior . . . . . . . . . . . . . . . . . . 11 76 6.1. Binding Table Maintenance . . . . . . . . . . . . . . . . 11 77 6.2. lwAFTR Data Plane Behavior . . . . . . . . . . . . . . . 12 78 7. Additional IPv4 address and Port Set Provisioning 79 Mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . 13 80 8. ICMP Processing . . . . . . . . . . . . . . . . . . . . . . . 14 81 8.1. ICMPv4 Processing by the lwAFTR . . . . . . . . . . . . . 14 82 8.2. ICMPv4 Processing by the lwB4 . . . . . . . . . . . . . . 14 83 9. Security Considerations . . . . . . . . . . . . . . . . . . . 15 84 10. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 15 85 11. Author List . . . . . . . . . . . . . . . . . . . . . . . . . 15 86 12. Acknowledgement . . . . . . . . . . . . . . . . . . . . . . . 19 87 13. References . . . . . . . . . . . . . . . . . . . . . . . . . 19 88 13.1. Normative References . . . . . . . . . . . . . . . . . . 19 89 13.2. Informative References . . . . . . . . . . . . . . . . . 20 90 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 21 92 1. Introduction 94 Dual-Stack Lite (DS-Lite, [RFC6333]) defines a model for providing 95 IPv4 access over an IPv6 network using two well-known technologies: 96 IP in IP [RFC2473] and Network Address Translation (NAT). The DS- 97 Lite architecture defines two major functional elements as follows: 99 Basic Bridging BroadBand element: A B4 element is a function 100 implemented on a dual-stack capable 101 node, either a directly connected 102 device or a CPE, that creates an 103 IPv4-in-IPv6 tunnel to an AFTR. 105 Address Family Transition Router: An AFTR element is the combination 106 of an IPv4-in-IPv6 tunnel endpoint 107 and an IPv4-IPv4 NAT implemented on 108 the same node. 110 As the AFTR performs the centralized NAT44 function, it dynamically 111 assigns public IPv4 addresses and ports to requesting host's traffic 112 (as described in [RFC3022]). To achieve this, the AFTR must 113 dynamically maintain per-flow state in the form of active NAPT 114 sessions. For service providers with a large number of B4 clients, 115 the size and associated costs for scaling the AFTR can quickly become 116 prohibitive. It can also place a large NAPT logging overhead upon 117 the service provider in countries where legal requirements mandate 118 this. 120 This document describes a mechanism called Lightweight 4 over 6 121 (lw4o6), which provides a solution for these problems. By relocating 122 the NAPT functionality from the centralized AFTR to the distributed 123 B4s, a number of benefits can be realised: 125 o NAPT44 functionality is already widely supported and used in 126 today's CPE devices. Lw4o6 uses this to provide private<->public 127 NAPT44, meaning that the service provider does not need a 128 centralized NAT44 function. 130 o The amount of state that must be maintained centrally in the AFTR 131 can be reduced from per-flow to per-subscriber. This reduces the 132 amount of resources (memory and processing power) necessary in the 133 AFTR. 135 o The reduction of maintained state results in a greatly reduced 136 logging overhead on the service provider. 138 Operator's IPv6 and IPv4 addressing architectures remain independent 139 of each other. Therefore, flexible IPv4/IPv6 addressing schemes can 140 be deployed. 142 Lightweight 4over6 is a solution designed specifically for complete 143 independence between IPv6 subnet prefix and IPv4 address with or 144 without IPv4 address sharing. This is accomplished by maintaining 145 state for each softwire (per-subscriber state) in the central lwAFTR 146 and a hub-and-spoke forwarding architecture. [I-D.ietf-softwire-map] 147 also offers these capabilities or, alternatively, allows for a 148 reduction of the amount of centralized state using rules to express 149 IPv4/IPv6 address mappings. This introduces an algorithmic 150 relationship between the IPv6 subnet and IPv4 address. This 151 relationship also allows the option of direct, meshed connectivity 152 between users. 154 The tunneling mechanism remains the same for DS-Lite and Lightweight 155 4over6. This document describes the changes to DS-Lite that are 156 necessary to implement Lightweight 4over6. These changes mainly 157 concern the configuration parameters and provisioning method 158 necessary for the functional elements. 160 Lightweight 4over6 features keeping per-subscriber state in the 161 service provider's network. It is categorized as Binding approach in 162 [I-D.ietf-softwire-unified-cpe] which defines a unified IPv4-in-IPv6 163 Softwire CPE. 165 This document extends the mechanism defined in [RFC7040] by allowing 166 address sharing. The solution in this document is also a variant of 167 A+P called Binding Table Mode (see Section 4.4 of [RFC6346]). 169 This document focuses on architectural considerations and 170 particularly on the expected behavior of the involved functional 171 elements and their interfaces. Deployment-specific issues are 172 discussed in a companion document. As such, discussions about 173 redundancy and provisioning policy are out of scope. 175 2. Conventions 177 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 178 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 179 document are to be interpreted as described in [RFC2119]. 181 3. Terminology 183 The document defines the following terms: 185 Lightweight 4over6 (lw4o6): An IPv4-over-IPv6 hub and spoke 186 mechanism, which extends DS-Lite by 187 moving the IPv4 translation (NAPT44) 188 function from the AFTR to the B4. 190 Lightweight B4 (lwB4): A B4 element (Basic Bridging BroadBand 191 element [RFC6333]), which supports 192 Lightweight 4over6 extensions. An lwB4 193 is a function implemented on a dual- 194 stack capable node, (either a directly 195 connected device or a CPE), that 196 supports port-restricted IPv4 address 197 allocation, implements NAPT44 198 functionality and creates a tunnel to 199 an lwAFTR. 201 Lightweight AFTR (lwAFTR): An AFTR element (Address Family 202 Transition Router element [RFC6333]), 203 which supports Lightweight 4over6 204 extension. An lwAFTR is an IPv4-in- 205 IPv6 tunnel endpoint which maintains 206 per-subscriber address binding only and 207 does not perform a NAPT44 function. 209 Restricted Port-Set: A non-overlapping range of allowed 210 external ports allocated to the lwB4 to 211 use for NAPT44. Source ports of IPv4 212 packets sent by the B4 must belong to 213 the assigned port-set. The port set is 214 used for all port aware IP protocols 215 (TCP, UDP, SCTP etc.). 217 Port-restricted IPv4 Address: A public IPv4 address with a restricted 218 port-set. In Lightweight 4over6, 219 multiple B4s may share the same IPv4 220 address, however, their port-sets must 221 be non-overlapping. 223 Throughout the remainder of this document, the terms B4/AFTR should 224 be understood to refer specifically to a DS-Lite implementation. The 225 terms lwB4/lwAFTR refer to a Lightweight 4over6 implementation. 227 4. Lightweight 4over6 Architecture 229 The Lightweight 4over6 architecture is functionally similar to DS- 230 Lite. lwB4s and an lwAFTR are connected through an IPv6-enabled 231 network. Both approaches use an IPv4-in-IPv6 encapsulation scheme to 232 deliver IPv4 connectivity. The following figure shows the data plane 233 with the main functional change between DS-Lite and lw4o6: 235 +--------+ +---------+ IPv4-in-IPv6 +---------+ +-------------+ 236 |IPv4 LAN|---| B4 |================|AFTR/NAPT|----|IPv4 Internet| 237 +--------+ +---------+ +---------+ +-------------+ 238 DS-Lite NAPT model: all state in the AFTR 240 +--------+ +---------+ IPv4-in-IPv6 +------+ +-------------+ 241 |IPv4 LAN|---|lwB4/NAPT|================|lwAFTR|----|IPv4 Internet| 242 +--------+ +---------+ +------+ +-------------+ 243 LW4over6 NAPT model: 244 subscriber state in the lwAFTR, NAPT state in lwB4 246 Figure 1 Comparison of DS-Lite and Lightweight 4over6 Data Plane 248 There are three main components in the Lightweight 4over6 249 architecture: 251 o The lwB4, which performs the NAPT function and encapsulation/de- 252 capsulation IPv4/IPv6. 254 o The lwAFTR, which performs the encapsulation/de-capsulation IPv4/ 255 IPv6. 257 o The provisioning system, which tells the lwB4 which IPv4 address 258 and port set to use. 260 The lwB4 differs from a regular B4 in that it now performs the NAPT 261 functionality. This means that it needs to be provisioned with the 262 public IPv4 address and port set it is allowed to use. This 263 information is provided though a provisioning mechanism such as DHCP, 264 Port Control Protocol (PCP, [RFC6887]) or TR-69. 266 The lwAFTR needs to know the binding between the IPv6 address of each 267 subscriber and the IPv4 address and port set allocated to that 268 subscriber. This information is used to perform ingress filtering 269 upstream and encapsulation downstream. Note that this is per- 270 subscriber state as opposed to per-flow state in the regular AFTR 271 case. 273 The consequence of this architecture is that the information 274 maintained by the provisioning mechanism and the one maintained by 275 the lwAFTR MUST be synchronized (See figure 2). The precise 276 mechanism whereby this synchronization occurs is out of scope for 277 this document. 279 The solution specified in this document allows the assignment of 280 either a full or a shared IPv4 address to requesting CPEs. [RFC7040] 281 provides a mechanism for assigning a full IPv4 address only. 283 +------------+ 284 /-------|Provisioning|<-----\ 285 | +------------+ | 286 | | 287 V V 288 +--------+ +---------+ IPv4/IPv6 +------+ +-------------+ 289 |IPv4 LAN|---|lwB4/NAPT|==================|lwAFTR|----|IPv4 Internet| 290 +--------+ +---------+ +------+ +-------------+ 292 Figure 2 Lightweight 4over6 Provisioning Synchronization 294 5. Lightweight B4 Behavior 296 5.1. Lightweight B4 Provisioning with DHCPv6 298 With DS-Lite, the B4 element only needs to be configured with a 299 single DS-Lite specific parameter so that it can set up the softwire 300 (the IPv6 address of the AFTR). Its IPv4 address can be taken from 301 the well-known range 192.0.0.0/29. 303 In lw4o6, a number of lw4o6 specific configuration parameters must be 304 provisioned to the lwB4. These are: 306 o IPv6 Address for the lwAFTR 308 o IPv4 External (Public) Address for NAPT44 310 o Restricted port-set to use for NAPT44 312 o IPv6 Binding Prefix 314 The lwB4 MUST implement DHCPv6 based configuration using 315 OPTION_S46_CONT_LW as described in section 5.3 of 316 [I-D.ietf-softwire-map-dhcp]. This means that the lifetime of the 317 softwire and the derived configuration information (e.g. IPv4 shared 318 address, IPv4 address) is bound to the lifetime of the DHCPv6 lease. 319 If stateful IPv4 configuration or additional IPv4 configuration 320 information is required, DHCP 4o6 [RFC7341] MUST be used. 322 Although it would be possible to extend lw4o6 to have more than one 323 active lw4o6 tunnel configured simultaneously, this document is only 324 concerned with the use of a single tunnel. 326 The IPv6 binding prefix field is provisioned so that the CE can 327 identify the correct prefix to use as the tunnel source. On receipt 328 of the necessary configuration parameters listed above, the lwB4 329 performs a longest prefix match between the IPv6 binding prefix and 330 its currently active IPv6 prefixes. The result forms the subnet to 331 be used for sourcing the lw4o6 tunnel. The full /128 address is then 332 constructed in the same manner as [I-D.ietf-softwire-map]. 334 0 1 2 3 335 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 1 336 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 337 | Operator Assigned Prefix | 338 . (64-bits) . 339 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 340 | Zero Padding | IPv4 Address | 341 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 342 | IPv4 Addr cont. | PSID | 343 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 345 Figure 3 Construction of the lw4o6 /128 Prefix 347 Operator Assigned Prefix: IPv6 prefix allocated to the client. If 348 the prefix length is less than 64, right padded with 349 zeros to 64-bits. 351 Padding: Padding (all zeros) 353 IPv4 Address: Public IPv4 address allocated to the client 355 PSID: Port Set ID allocated to the client, left padded with 356 zeros to 16-bits. If no PSID is provisioned, all 357 zeros. 359 In the event that the lwB4's IPv6 encapsulation source address is 360 changed for any reason (such as the DHCPv6 lease expiring), the 361 lwB4's dynamic provisioning process MUST be re-initiated. When the 362 lwB4's public IPv4 address or port set ID is changed for any reason, 363 the lwB4 MUST flush its NAPT table. 365 An lwB4 MUST support dynamic port-restricted IPv4 address 366 provisioning. The port set algorithm for provisioning this is 367 described in Section 5.1 of [I-D.ietf-softwire-map]. For lw4o6, the 368 number of a-bits SHOULD be 0, thus allocating a single contiguous 369 port set to each lwB4. 371 Provisioning of the lwB4 using DHCPv6 as described here allocates a 372 single PSID to the client. In the event that the client is 373 concurrently using all of the provisioned L4 ports it may be unable 374 to initiate any additional outbound connections. DHCPv6 based 375 provisioning does not provide a mechanism for the client to request 376 more L4 port numbers. Other provisioning mechanisms (e.g. PCP based 377 provisioning [I-D.ietf-pcp-port-set]) provide this function. Issues 378 relevant to IP address sharing are discussed in more detail in 379 [RFC6269]. 381 Unless an lwB4 is being allocated a full IPv4 address, it is 382 RECOMMENDED that PSIDs containing the system ports (0-1023) are not 383 allocated to lwB4s. The reserved ports are more likely to be 384 reserved by middleware, and therefore we recommend that they not be 385 issued to clients other than as a deliberate assignment. 386 Section 5.2.2 of [RFC6269] provides analysis of allocating system 387 ports to clients with IPv4 address sharing. 389 In the event that the lwB4 receives an ICMPv6 error message (type 1, 390 code 5) originating from the lwAFTR, the lwB4 interprets this to mean 391 that no matching entry in the lwAFTR's binding table has been found, 392 so the IPv4 payload is not being forwarded by the lwAFTR. The lwB4 393 MAY then re-initiate the dynamic port-restricted provisioning 394 process. The lwB4's re-initiation policy SHOULD be configurable. 396 On receipt of such an ICMP error message, the lwB4 MUST validate the 397 source address to be the same as the lwAFTR address that is 398 configured. In the event that these addresses do not match, the 399 lwAFTR MUST discard the ICMP error message. 401 In order to prevent forged ICMP messages (using the spoofed lwAFTR 402 address as the source) from being sent to lwB4s, the operator can 403 implement network ingress filtering as described in [RFC2827]. 405 The DNS considerations described in Section 5.5 and Section 6.4 of 406 [RFC6333] apply to Lightweight 4over6; lw4o6 implementations MUST 407 comply with all requirements stated there. 409 5.2. Lightweight B4 Data Plane Behavior 411 Several sections of [RFC6333] provide background information on the 412 B4's data plane functionality and MUST be implemented by the lwB4 as 413 they are common to both solutions. The relevant sections are: 415 5.2 Encapsulation Covering encapsulation and de- 416 capsulation of tunneled traffic 418 5.3 Fragmentation and Reassembly Covering MTU and fragmentation 419 considerations (referencing 420 [RFC2473]). 422 7.1 Tunneling Covering tunneling and traffic 423 class mapping between IPv4 and IPv6 424 (referencing [RFC2473] and 425 [RFC2983]) 427 The lwB4 element performs IPv4 address translation (NAPT44) as well 428 as encapsulation and de-capsulation. It runs standard NAPT44 429 [RFC3022] using the allocated port-restricted address as its external 430 IPv4 address and port numbers. 432 The working flow of the lwB4 is illustrated in figure 4. 434 +-------------+ 435 | lwB4 | 436 +--------+ IPv4 |------+------| IPv4-in-IPv6 +----------+ 437 |IPv4 LAN|------->| |Encap.|-------------->|Configured| 438 | |<-------| NAPT | or |<--------------| lwAFTR | 439 +--------+ | |Decap.| +----------+ 440 +------+------+ 442 Figure 4 Working Flow of the lwB4 444 Hosts connected to the customer's network behind the lwB4 source IPv4 445 packets with an [RFC1918] address. When the lwB4 receives such an 446 IPv4 packet, it performs a NAPT44 function on the source address and 447 port by using the public IPv4 address and a port number from the 448 allocated port-set. Then, it encapsulates the packet with an IPv6 449 header. The destination IPv6 address is the lwAFTR's IPv6 address 450 and the source IPv6 address is the lwB4's IPv6 tunnel endpoint 451 address. Finally, the lwB4 forwards the encapsulated packet to the 452 configured lwAFTR. 454 When the lwB4 receives an IPv4-in-IPv6 packet from the lwAFTR, it de- 455 capsulates the IPv4 packet from the IPv6 packet. Then, it performs 456 NAPT44 translation on the destination address and port, based on the 457 available information in its local NAPT44 table. 459 If the IPv6 source address does not match the configured lwAFTR 460 address, then the packet MUST be discarded. If the decapsulated IPv4 461 packet does not match the lwB4's configuration (i.e. invalid 462 destination IPv4 address or port) then the packet MUST be dropped. 463 An ICMPv4 error message (type 13 - Communication Administratively 464 Prohibited) message MAY be sent back to the lwAFTR. The ICMP policy 465 SHOULD be configurable. 467 The lwB4 is responsible for performing ALG functions (e.g., SIP, 468 FTP), and other NAPT traversal mechanisms (e.g., UPnP, NAPT-PMP, 469 manual binding configuration, PCP) for the internal hosts, if 470 necessary. This requirement is typical for NAPT44 gateways available 471 today. 473 It is possible that a lwB4 is co-located in a host. In this case, 474 the functions of NAPT44 and encapsulation/de-capsulation are 475 implemented inside the host. 477 5.2.1. Fragmentation Behaviour 479 For TCP and UDP traffic the NAPT44 implemented in the lwB4 MUST 480 conform with the behaviour and best current practices documented in 481 [RFC4787], [RFC5508], and [RFC5382]. If the lwB4 supports DCCP, then 482 the requirements in [RFC5597] MUST be implemented. 484 The NAPT44 in the lwB4 MUST implement ICMP message handling behaviour 485 conforming to the best current practice documented in [RFC5508]. If 486 the lwB4 receives an ICMP error (for errors detected inside the IPv6 487 tunnel), the node relays the ICMP error message to the original 488 source (the lwAFTR). This behaviour SHOULD be implemented conforming 489 to the section 8 of [RFC2473]. 491 If IPv4 hosts behind different lwB4s sharing the same IPv4 address 492 send fragments to the same IPv4 destination host outside the 493 Lightweight 4over6 domain, those hosts may use the same IPv4 494 fragmentation identifier, resulting in incorrect reassembly of the 495 fragments at the destination host. Given that the IPv4 fragmentation 496 identifier is a 16-bit field, it could be used similarly to port 497 ranges: A lwB4 could rewrite the IPv4 fragmentation identifier to be 498 within its allocated port-set, if the resulting fragment identifier 499 space is large enough related to the rate fragments are sent. 500 However, splitting the identifier space in this fashion would 501 increase the probability of reassembly collision for all connections 502 through the lwB4. See also Section 5.3.1 of [RFC6864]. 504 6. Lightweight AFTR Behavior 506 6.1. Binding Table Maintenance 508 The lwAFTR maintains an address binding table containing the binding 509 between the lwB4's IPv6 address, the allocated IPv4 address and 510 restricted port-set. Unlike the DS-Lite extended binding table 511 defined in section 6.6 of [RFC6333] which is a 5-tuple NAPT table, 512 each entry in the Lightweight 4over6 binding table contains the 513 following 3-tuples: 515 o IPv6 Address for a single lwB4 517 o Public IPv4 Address 519 o Restricted port-set 520 The entry has two functions: the IPv6 encapsulation of inbound IPv4 521 packets destined to the lwB4 and the validation of outbound IPv4-in- 522 IPv6 packets received from the lwB4 for de-capsulation. 524 The lwAFTR does not perform NAPT and so does not need session 525 entries. 527 The lwAFTR MUST synchronize the binding information with the port- 528 restricted address provisioning process. If the lwAFTR does not 529 participate in the port-restricted address provisioning process, the 530 binding MUST be synchronized through other methods (e.g. out-of-band 531 static update). 533 If the lwAFTR participates in the port-restricted provisioning 534 process, then its binding table MUST be created as part of this 535 process. 537 For all provisioning processes, the lifetime of binding table entries 538 MUST be synchronized with the lifetime of address allocations. 540 6.2. lwAFTR Data Plane Behavior 542 Several sections of [RFC6333] provide background information on the 543 AFTR's data plane functionality and MUST be implemented by the lwAFTR 544 as they are common to both solutions. The relevant sections are: 546 6.2 Encapsulation Covering encapsulation and de- 547 capsulation of tunneled traffic 549 6.3 Fragmentation and Reassembly Fragmentation and re-assembly 550 considerations (referencing 551 [RFC2473]) 553 7.1 Tunneling Covering tunneling and traffic 554 class mapping between IPv4 and IPv6 555 (referencing [RFC2473] and 556 [RFC2983]) 558 When the lwAFTR receives an IPv4-in-IPv6 packet from an lwB4, it de- 559 capsulates the IPv6 header and verifies the source addresses and port 560 in the binding table. If both the source IPv4 and IPv6 addresses 561 match a single entry in the binding table and the source port is in 562 the allowed port-set for that entry, the lwAFTR forwards the packet 563 to the IPv4 destination. 565 If no match is found (e.g., no matching IPv4 address entry, port out 566 of range, etc.), the lwAFTR MUST discard or implement a policy (such 567 as redirection) on the packet. An ICMPv6 type 1, code 5 (source 568 address failed ingress/egress policy) error message MAY be sent back 569 to the requesting lwB4. The ICMP policy SHOULD be configurable. 571 When the lwAFTR receives an inbound IPv4 packet, it uses the IPv4 572 destination address and port to lookup the destination lwB4's IPv6 573 address in its binding table. If a match is found, the lwAFTR 574 encapsulates the IPv4 packet. The source is the lwAFTR's IPv6 575 address and the destination is the lwB4's IPv6 address from the 576 matched entry. Then, the lwAFTR forwards the packet to the lwB4 577 natively over the IPv6 network. 579 If no match is found, the lwAFTR MUST discard the packet. An ICMPv4 580 type 3, code 1 (Destination unreachable, host unreachable) error 581 message MAY be sent back. The ICMP policy SHOULD be configurable. 583 The lwAFTR MUST support hairpinning of traffic between two lwB4s, by 584 performing de-capsulation and re-encapsulation of packets from one 585 lwB4 that need to be sent to another lwB4 associated with the same 586 AFTR. The hairpinning policy MUST be configurable. 588 7. Additional IPv4 address and Port Set Provisioning Mechanisms 590 In addition to the DHCPv6 based mechanism described in section 5.1, 591 several other IPv4 provisioning protocols have been suggested. These 592 protocols MAY be implemented. These alternatives include: 594 o DHCPv4 over DHCPv6: [RFC7341] describes implementing DHCPv4 595 messages over an IPv6 only service providers network. This 596 enables leasing of IPv4 addresses and makes DHCPv4 options 597 available to the DHCPv4-over-DHCPv6 client. An lwB4 MAY implement 598 [RFC7341] and [I-D.ietf-dhc-dynamic-shared-v4allocation] to 599 retrieve a shared IPv4 address with a set of ports. 601 o PCP[RFC6887]: an lwB4 MAY use [I-D.ietf-pcp-port-set] to retrieve 602 a restricted IPv4 address and a set of ports. 604 In a Lightweight 4over6 domain, the binding information MUST be 605 synchronized across the lwB4s, the lwAFTRs and the provisioning 606 server. 608 To prevent interworking complexity, it is RECOMMENDED that an 609 operator uses a single provisioning mechanism / protocol for their 610 implementation. In the event that more than one provisioning 611 mechanism / protocol needs to be used (for example during a migration 612 to a new provisioning mechanism), the operator SHOULD ensure that 613 each provisioning mechanism has a discrete set of resources (e.g. 615 IPv4 address/PSID pools and lwAFTR tunnel addresses and binding 616 tables). 618 8. ICMP Processing 620 For both the lwAFTR and the lwB4, ICMPv6 MUST be handled as described 621 in [RFC2473]. 623 ICMPv4 does not work in an address sharing environment without 624 special handling [RFC6269]. Due to the port-set style address 625 sharing, Lightweight 4over6 requires specific ICMP message handling 626 not required by DS-Lite. 628 8.1. ICMPv4 Processing by the lwAFTR 630 For inbound ICMP messages The following behavior SHOULD be 631 implemented by the lwAFTR to provide ICMP error handling and basic 632 remote IPv4 service diagnostics for a port restricted CPE: 634 1. Check the ICMP Type field. 636 2. If the ICMP type is set to 0 or 8 (echo reply or request), then 637 the lwAFTR MUST take the value of the ICMP identifier field as 638 the source port, and use this value to lookup the binding table 639 for an encapsulation destination. If a match is found, the 640 lwAFTR forwards the ICMP packet to the IPv6 address stored in the 641 entry; otherwise it MUST discard the packet. 643 3. If the ICMP type field is set to any other value, then the lwAFTR 644 MUST use the method described in REQ-3 of [RFC5508] to locate the 645 source port within the transport layer header in ICMP packet's 646 data field. The destination IPv4 address and source port 647 extracted from the ICMP packet are then used to make a lookup in 648 the binding table. If a match is found, it MUST forward the ICMP 649 reply packet to the IPv6 address stored in the entry; otherwise 650 it MUST discard the packet. 652 Otherwise the lwAFTR MUST discard all inbound ICMPv4 messages. 654 The ICMP policy SHOULD be configurable. 656 8.2. ICMPv4 Processing by the lwB4 658 The lwB4 MUST implement the requirements defined in [RFC5508] for 659 ICMP forwarding. For ICMP echo request packets originating from the 660 private IPv4 network, the lwB4 SHOULD implement the method described 661 in [RFC6346] and use an available port from its port-set as the ICMP 662 Identifier. 664 9. Security Considerations 666 As the port space for a subscriber shrinks due to address sharing, 667 the randomness for the port numbers of the subscriber is decreased 668 significantly. This means it is much easier for an attacker to guess 669 the port number used, which could result in attacks ranging from 670 throughput reduction to broken connections or data corruption. 672 The port-set for a subscriber can be a set of contiguous ports or 673 non-contiguous ports. Contiguous port-sets do not reduce this 674 threat. However, with non-contiguous port-set (which may be 675 generated in a pseudo-random way [RFC6431]), the randomness of the 676 port number is improved, provided that the attacker is outside the 677 Lightweight 4over6 domain and hence does not know the port-set 678 generation algorithm. 680 The lwAFTR MUST rate limit ICMPv6 error messages (see Section 5.1) to 681 defend against DoS attacks generated by an abuse user. 683 More considerations about IP address sharing are discussed in 684 Section 13 of [RFC6269], which is applicable to this solution. 686 This document describes a number of different protocols which may be 687 used for the provisioning of lw4o6. In each case, the security 688 considerations relevant to the provisioning protocol are also 689 relevant to the provisioning of lw4o6 using that protocol. Lw4o6 690 does not add any additional provisioning protocol specific security 691 considerations. 693 10. IANA Considerations 695 This document does not include an IANA request. 697 11. Author List 699 The following are extended authors who contributed to the effort: 701 Jianping Wu 703 Tsinghua University 705 Department of Computer Science, Tsinghua University 707 Beijing 100084 709 P.R.China 711 Phone: +86-10-62785983 712 Email: jianping@cernet.edu.cn 714 Peng Wu 716 Tsinghua University 718 Department of Computer Science, Tsinghua University 720 Beijing 100084 722 P.R.China 724 Phone: +86-10-62785822 726 Email: pengwu.thu@gmail.com 728 Qi Sun 730 Tsinghua University 732 Beijing 100084 734 P.R.China 736 Phone: +86-10-62785822 738 Email: sunqi@csnet1.cs.tsinghua.edu.cn 740 Chongfeng Xie 742 China Telecom 744 Room 708, No.118, Xizhimennei Street 746 Beijing 100035 748 P.R.China 750 Phone: +86-10-58552116 752 Email: xiechf@ctbri.com.cn 753 Xiaohong Deng 755 France Telecom 757 Email: xiaohong.deng@orange.com 759 Cathy Zhou 761 Huawei Technologies 763 Section B, Huawei Industrial Base, Bantian Longgang 765 Shenzhen 518129 767 P.R.China 769 Email: cathyzhou@huawei.com 771 Alain Durand 773 Juniper Networks 775 1194 North Mathilda Avenue 777 Sunnyvale, CA 94089-1206 779 USA 781 Email: adurand@juniper.net 783 Reinaldo Penno 785 Cisco Systems, Inc. 787 170 West Tasman Drive 789 San Jose, California 95134 791 USA 793 Email: repenno@cisco.com 794 Axel Clauberg 796 Deutsche Telekom AG 798 CTO-ATI 800 Landgrabenweg 151 802 Bonn, 53227 804 Germany 806 Email: axel.clauberg@telekom.de 808 Lionel Hoffmann 810 Bouygues Telecom 812 TECHNOPOLE 814 13/15 Avenue du Marechal Juin 816 Meudon 92360 818 France 820 Email: lhoffman@bouyguestelecom.fr 822 Maoke Chen 824 FreeBit Co., Ltd. 826 13F E-space Tower, Maruyama-cho 3-6 828 Shibuya-ku, Tokyo 150-0044 830 Japan 832 Email: fibrib@gmail.com 834 12. Acknowledgement 836 The authors would like to thank Ole Troan, Ralph Droms and Suresh 837 Krishnan for their comments and feedback. 839 This document is a merge of three documents: 840 [I-D.cui-softwire-b4-translated-ds-lite], [I-D.zhou-softwire-b4-nat] 841 and [I-D.penno-softwire-sdnat]. 843 13. References 845 13.1. Normative References 847 [I-D.ietf-softwire-map-dhcp] 848 Mrugalski, T., Troan, O., Farrer, I., Perreault, S., Dec, 849 W., Bao, C., leaf.yeh.sdo@gmail.com, l., and X. Deng, 850 "DHCPv6 Options for configuration of Softwire Address and 851 Port Mapped Clients", draft-ietf-softwire-map-dhcp-10 852 (work in progress), November 2014. 854 [RFC1918] Rekhter, Y., Moskowitz, R., Karrenberg, D., Groot, G., and 855 E. Lear, "Address Allocation for Private Internets", BCP 856 5, RFC 1918, February 1996. 858 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 859 Requirement Levels", BCP 14, RFC 2119, March 1997. 861 [RFC2473] Conta, A. and S. Deering, "Generic Packet Tunneling in 862 IPv6 Specification", RFC 2473, December 1998. 864 [RFC4787] Audet, F. and C. Jennings, "Network Address Translation 865 (NAT) Behavioral Requirements for Unicast UDP", BCP 127, 866 RFC 4787, January 2007. 868 [RFC5382] Guha, S., Biswas, K., Ford, B., Sivakumar, S., and P. 869 Srisuresh, "NAT Behavioral Requirements for TCP", BCP 142, 870 RFC 5382, October 2008. 872 [RFC5508] Srisuresh, P., Ford, B., Sivakumar, S., and S. Guha, "NAT 873 Behavioral Requirements for ICMP", BCP 148, RFC 5508, 874 April 2009. 876 [RFC5597] Denis-Courmont, R., "Network Address Translation (NAT) 877 Behavioral Requirements for the Datagram Congestion 878 Control Protocol", BCP 150, RFC 5597, September 2009. 880 [RFC6333] Durand, A., Droms, R., Woodyatt, J., and Y. Lee, "Dual- 881 Stack Lite Broadband Deployments Following IPv4 882 Exhaustion", RFC 6333, August 2011. 884 13.2. Informative References 886 [I-D.cui-softwire-b4-translated-ds-lite] 887 Cui, Y., Sun, Q., Boucadair, M., Tsou, T., Lee, Y., and I. 888 Farrer, "Lightweight 4over6: An Extension to the DS-Lite 889 Architecture", draft-cui-softwire-b4-translated-ds-lite-11 890 (work in progress), February 2013. 892 [I-D.ietf-dhc-dynamic-shared-v4allocation] 893 Cui, Y., Qiong, Q., Farrer, I., Lee, Y., Sun, Q., and M. 894 Boucadair, "Dynamic Allocation of Shared IPv4 Addresses", 895 draft-ietf-dhc-dynamic-shared-v4allocation-02 (work in 896 progress), September 2014. 898 [I-D.ietf-pcp-port-set] 899 Qiong, Q., Boucadair, M., Sivakumar, S., Zhou, C., Tsou, 900 T., and S. Perreault, "Port Control Protocol (PCP) 901 Extension for Port Set Allocation", draft-ietf-pcp-port- 902 set-07 (work in progress), November 2014. 904 [I-D.ietf-softwire-map] 905 Troan, O., Dec, W., Li, X., Bao, C., Matsushima, S., 906 Murakami, T., and T. Taylor, "Mapping of Address and Port 907 with Encapsulation (MAP)", draft-ietf-softwire-map-11 908 (work in progress), October 2014. 910 [I-D.ietf-softwire-unified-cpe] 911 Boucadair, M., Farrer, I., Perreault, S., and S. 912 Sivakumar, "Unified IPv4-in-IPv6 Softwire CPE", draft- 913 ietf-softwire-unified-cpe-01 (work in progress), May 2013. 915 [I-D.penno-softwire-sdnat] 916 Penno, R., Durand, A., Hoffmann, L., and A. Clauberg, 917 "Stateless DS-Lite", draft-penno-softwire-sdnat-02 (work 918 in progress), March 2012. 920 [I-D.zhou-softwire-b4-nat] 921 Zhou, C., Boucadair, M., and X. Deng, "NAT offload 922 extension to Dual-Stack lite", draft-zhou-softwire- 923 b4-nat-04 (work in progress), October 2011. 925 [RFC2827] Ferguson, P. and D. Senie, "Network Ingress Filtering: 926 Defeating Denial of Service Attacks which employ IP Source 927 Address Spoofing", BCP 38, RFC 2827, May 2000. 929 [RFC2983] Black, D., "Differentiated Services and Tunnels", RFC 930 2983, October 2000. 932 [RFC3022] Srisuresh, P. and K. Egevang, "Traditional IP Network 933 Address Translator (Traditional NAT)", RFC 3022, January 934 2001. 936 [RFC6269] Ford, M., Boucadair, M., Durand, A., Levis, P., and P. 937 Roberts, "Issues with IP Address Sharing", RFC 6269, June 938 2011. 940 [RFC6346] Bush, R., "The Address plus Port (A+P) Approach to the 941 IPv4 Address Shortage", RFC 6346, August 2011. 943 [RFC6431] Boucadair, M., Levis, P., Bajko, G., Savolainen, T., and 944 T. Tsou, "Huawei Port Range Configuration Options for PPP 945 IP Control Protocol (IPCP)", RFC 6431, November 2011. 947 [RFC6864] Touch, J., "Updated Specification of the IPv4 ID Field", 948 RFC 6864, February 2013. 950 [RFC6887] Wing, D., Cheshire, S., Boucadair, M., Penno, R., and P. 951 Selkirk, "Port Control Protocol (PCP)", RFC 6887, April 952 2013. 954 [RFC7040] Cui, Y., Wu, J., Wu, P., Vautrin, O., and Y. Lee, "Public 955 IPv4-over-IPv6 Access Network", RFC 7040, November 2013. 957 [RFC7341] Sun, Q., Cui, Y., Siodelski, M., Krishnan, S., and I. 958 Farrer, "DHCPv4-over-DHCPv6 (DHCP 4o6) Transport", RFC 959 7341, August 2014. 961 Authors' Addresses 963 Yong Cui 964 Tsinghua University 965 Beijing 100084 966 P.R.China 968 Phone: +86-10-62603059 969 Email: yong@csnet1.cs.tsinghua.edu.cn 970 Qiong Sun 971 China Telecom 972 Room 708, No.118, Xizhimennei Street 973 Beijing 100035 974 P.R.China 976 Phone: +86-10-58552936 977 Email: sunqiong@ctbri.com.cn 979 Mohamed Boucadair 980 France Telecom 981 Rennes 35000 982 France 984 Email: mohamed.boucadair@orange.com 986 Tina Tsou 987 Huawei Technologies 988 2330 Central Expressway 989 Santa Clara, CA 95050 990 USA 992 Phone: +1-408-330-4424 993 Email: tena@huawei.com 995 Yiu L. Lee 996 Comcast 997 One Comcast Center 998 Philadelphia, PA 19103 999 USA 1001 Email: yiu_lee@cable.comcast.com 1003 Ian Farrer 1004 Deutsche Telekom AG 1005 CTO-ATI, Landgrabenweg 151 1006 Bonn, NRW 53227 1007 Germany 1009 Email: ian.farrer@telekom.de