idnits 2.17.1 draft-donley-behave-deterministic-cgn-07.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- == There are 18 instances 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 (January 13, 2014) is 3755 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- == Outdated reference: A later version (-06) exists of draft-donley-nat444-impacts-04 == Outdated reference: A later version (-06) exists of draft-sivakumar-behave-nat-logging-05 == Outdated reference: A later version (-06) exists of draft-tsou-behave-natx4-log-reduction-02 Summary: 0 errors (**), 0 flaws (~~), 5 warnings (==), 1 comment (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group C. Donley 3 Internet-Draft CableLabs 4 Intended status: Informational C. Grundemann 5 Expires: July 17, 2014 Internet Society 6 V. Sarawat 7 K. Sundaresan 8 CableLabs 9 O. Vautrin 10 Juniper Networks 11 January 13, 2014 13 Deterministic Address Mapping to Reduce Logging in Carrier Grade NAT 14 Deployments 15 draft-donley-behave-deterministic-cgn-07 17 Abstract 19 In some instances, Service Providers have a legal logging requirement 20 to be able to map a subscriber's inside address with the address used 21 on the public Internet (e.g. for abuse response). Unfortunately, 22 many Carrier Grade NAT logging solutions require active logging of 23 dynamic translations. Carrier Grade NAT port assignments are often 24 per-connection, but could optionally use port ranges. Research 25 indicates that per-connection logging is not scalable in many 26 residential broadband services. This document suggests a way to 27 manage Carrier Grade NAT translations in such a way as to 28 significantly reduce the amount of logging required while providing 29 traceability for abuse response. While the authors acknowledge that 30 IPv6 is a preferred solution, Carrier Grade NAT is a reality in many 31 networks, and is needed in situations where either customer equipment 32 or Internet content only supports IPv4; this approach should in no 33 way slow the deployment of IPv6. 35 Requirements Language 37 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 38 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 39 document are to be interpreted as described in RFC 2119 [RFC2119]. 41 Status of This Memo 43 This Internet-Draft is submitted in full conformance with the 44 provisions of BCP 78 and BCP 79. 46 Internet-Drafts are working documents of the Internet Engineering 47 Task Force (IETF). Note that other groups may also distribute 48 working documents as Internet-Drafts. The list of current Internet- 49 Drafts is at http://datatracker.ietf.org/drafts/current/. 51 Internet-Drafts are draft documents valid for a maximum of six months 52 and may be updated, replaced, or obsoleted by other documents at any 53 time. It is inappropriate to use Internet-Drafts as reference 54 material or to cite them other than as "work in progress." 56 This Internet-Draft will expire on July 17, 2014. 58 Copyright Notice 60 Copyright (c) 2014 IETF Trust and the persons identified as the 61 document authors. All rights reserved. 63 This document is subject to BCP 78 and the IETF Trust's Legal 64 Provisions Relating to IETF Documents 65 (http://trustee.ietf.org/license-info) in effect on the date of 66 publication of this document. Please review these documents 67 carefully, as they describe your rights and restrictions with respect 68 to this document. Code Components extracted from this document must 69 include Simplified BSD License text as described in Section 4.e of 70 the Trust Legal Provisions and are provided without warranty as 71 described in the Simplified BSD License. 73 Table of Contents 75 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 2 76 2. Deterministic Port Ranges . . . . . . . . . . . . . . . . . . 4 77 2.1. IPv4 Port Utilization Efficiency . . . . . . . . . . . . 7 78 2.2. Planning & Dimensioning . . . . . . . . . . . . . . . . . 8 79 2.3. Deterministic CGN Example . . . . . . . . . . . . . . . . 8 80 3. Additional Logging Considerations . . . . . . . . . . . . . . 10 81 3.1. Failover Considerations . . . . . . . . . . . . . . . . . 10 82 4. Impact on the IPv6 Transition . . . . . . . . . . . . . . . . 11 83 5. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 11 84 6. Security Considerations . . . . . . . . . . . . . . . . . . . 11 85 7. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 11 86 8. References . . . . . . . . . . . . . . . . . . . . . . . . . 12 87 8.1. Normative References . . . . . . . . . . . . . . . . . . 12 88 8.2. Informative References . . . . . . . . . . . . . . . . . 12 89 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 14 91 1. Introduction 93 It is becoming increasingly difficult to obtain new IPv4 address 94 assignments from Regional/Local Internet Registries due to depleting 95 supplies of unallocated IPv4 address space. To meet the growing 96 demand for Internet connectivity from new subscribers, devices, and 97 service types, some operators will be forced to share a single public 98 IPv4 address among multiple subscribers using techniques such as 99 Carrier Grade Network Address Translation (CGN) [RFC6264] (e.g., 100 NAT444 [I-D.shirasaki-nat444], DS-Lite [RFC6333], NAT64 [RFC6146] 101 etc.). However, address sharing poses additional challenges to 102 operators when considering how they manage service entitlement, 103 public safety requests, or attack/abuse/fraud reports [RFC6269]. In 104 order to identify a specific user associated with an IP address in 105 response to such a request or for service entitlement, an operator 106 will need to map a subscriber's internal source IP address and source 107 port with the global public IP address and source port provided by 108 the CGN for every connection initiated by the user. 110 CGN connection logging satisfies the need to identify attackers and 111 respond to abuse/public safety requests, but it imposes significant 112 operational challenges to operators. In lab testing, we have 113 observed CGN log messages to be approximately 150 bytes long for 114 NAT444 [I-D.shirasaki-nat444], and 175 bytes for DS-Lite [RFC6333] 115 (individual log messages vary somewhat in size). Although we are not 116 aware of definitive studies of connection rates per subscriber, 117 reports from several operators in the US sets the average number of 118 connections per household at approximately 33,000 connections per 119 day. If each connection is individually logged, this translates to a 120 data volume of approximately 5 MB per subscriber per day, or about 121 150 MB per subscriber per month; however, specific data volumes may 122 vary across different operators based on myriad factors. Based on 123 available data, a 1-million subscriber service provider will generate 124 approximately 150 terabytes of log data per month, or 1.8 petabytes 125 per year. 127 The volume of log data poses a problem for both operators and the 128 public safety community. On the operator side, it requires a 129 significant infrastructure investment by operators implementing CGN. 130 It also requires updated operational practices to maintain the 131 logging infrastructure, and requires approximately 23 Mbps of 132 bandwidth between the CGN devices and the logging infrastructure per 133 50,000 users. On the public safety side, it increases the time 134 required for an operator to search the logs in response to an abuse 135 report, and could delay investigations. Accordingly, an 136 international group of operators and public safety officials 137 approached the authors to identify a way to reduce this impact while 138 improving abuse response. 140 The volume of CGN logging can be reduced by assigning port ranges 141 instead of individual ports. Using this method, only the assignment 142 of a new port range is logged. This may massively reduce logging 143 volume. The log reduction may vary depending on the length of the 144 assigned port range, whether the port range is static or dynamic, 145 etc. This has been acknowledged in [RFC6269] and 146 [I-D.sivakumar-behave-nat-logging]. Per [RFC6269]: 148 "Address sharing solutions may mitigate these issues to some extent 149 by pre-allocating groups of ports. Then only the allocation of the 150 group needs to be recorded, and not the creation of every session 151 binding within that group. There are trade-offs to be made between 152 the sizes of these port groups, the ratio of public addresses to 153 subscribers, whether or not these groups timeout, and the impact on 154 logging requirements and port randomization security (RFC6056) 155 [RFC6056]." 157 However, the existing solution still poses an impact on operators and 158 public safety officials for logging and searching. Instead, CGNs 159 could be designed and/or configured to deterministically map internal 160 addresses to {external address + port range} in such a way as to be 161 able to algorithmically calculate the mapping. Only inputs and 162 configuration of the algorithm need to be logged. This approach 163 reduces both logging volume and subscriber identification times. In 164 some cases, when full deterministic allocation is used, this approach 165 can eliminate the need for translation logging. 167 This document describes a method for such CGN address mapping, 168 combined with block port reservations, that significantly reduces the 169 burden on operators while offering the ability to map a subscriber's 170 inside IP address with an outside address and external port number 171 observed on the Internet. 173 The activation of the proposed port range allocation scheme is 174 compliant with BEHAVE requirements such as the support of APP. 176 2. Deterministic Port Ranges 178 While a subscriber uses thousands of connections per day, most 179 subscribers use far fewer resources at any given time. When the 180 compression ratio (see Appendix B of RFC6269 [RFC6269]) is low (e.g., 181 the ratio of the number of subscribers to the number of public IPv4 182 addresses allocated to a CGN is closer to 10:1 than 1000:1), each 183 subscriber could expect to have access to thousands of TCP/UDP ports 184 at any given time. Thus, as an alternative to logging each 185 connection, CGNs could deterministically map customer private 186 addresses (received on the customer-facing interface of the CGN, 187 a.k.a., internal side) to public addresses extended with port ranges 188 (used on the Internet-facing interface of the CGN, a.k.a., external 189 side). This algorithm allows an operator to identify a subscriber 190 internal IP address when provided the public side IP and port number 191 without having to examine the CGN translation logs. This prevents an 192 operator from having to transport and store massive amounts of 193 session data from the CGN and then process it to identify a 194 subscriber. 196 The algorithmic mapping can be expressed as: 198 (External IP Address, Port Range) = function 1 (Internal IP Address) 200 Internal IP Address = function 2 (External IP Address, Port Number) 202 The CGN SHOULD provide a method for administrators to test both 203 mapping functions (e.g., enter an External IP Address + Port Number 204 and receive the corresponding Internal IP Address). 206 Deterministic Port Range allocation requires configuration of the 207 following variables: 209 o Inside IPv4/IPv6 address range (I); 211 o Outside IPv4 address range (O); 213 o Compression ratio (e.g. inside IP addresses I/outside IP addresses 214 O) (C); 216 o Dynamic address pool factor (D), to be added to the compression 217 ratio in order to create an overflow address pool; 219 o Maximum ports per user (M); 221 o Address assignment algorithm (A) (see below); and 223 o Reserved TCP/UDP port list (R) 225 Note: The inside address range (I) will be an IPv4 range in NAT444 226 operation (NAT444 [I-D.shirasaki-nat444]) and an IPv6 range in DS- 227 Lite operation (DS-Lite [RFC6333]). 229 A subscriber is identified by an internal IPv4 address (e.g., NAT44) 230 or an IPv6 prefix (e.g., DS-Lite or NAT64). 232 The algorithm may be generalized to L2-aware NAT 233 [I-D.miles-behave-l2nat] but this requires the configuration of the 234 Internal interface identifiers (e.g., MAC addresses). 236 The algorithm is not designed to retrieve an internal host among 237 those sharing the same internal IP address (e.g., in a DS-Lite 238 context, only an IPv6 address/prefix can be retrieved using the 239 algorithm while the internal IPv4 address used for the encapsulated 240 IPv4 datagram is lost). 242 Several address assignment algorithms are possible. Using predefined 243 algorithms, such as those that follow, simplifies the process of 244 reversing the algorithm when needed. However, the CGN MAY support 245 additional algorithms. Also, the CGN is not required to support all 246 algorithms described below. Subscribers could be restricted to ports 247 from a single IPv4 address, or could be allocated ports across all 248 addresses in a pool, for example. The following algorithms and 249 corresponding values of A are as follow: 251 0: Sequential (e.g. the first block goes to address 1, the second 252 block to address 2, etc.) 254 1: Staggered (e.g. for every n between 0 and ((65536-R)/(C+D))-1 , 255 address 1 receives ports n*C+R, address 2 receives ports 256 (1+n)*C+R, etc.) 258 2: Round robin (e.g. the subscriber receives the same port number 259 across a pool of external IP addresses. If the subscriber is to 260 be assigned more ports than there are in the external IP pool, the 261 subscriber receives the next highest port across the IP pool, and 262 so on. Thus, if there are 10 IP addresses in a pool and a 263 subscriber is assigned 1000 ports, the subscriber would receive a 264 range such as ports 2000-2099 across all 10 external IP 265 addresses). 267 3: Interlaced horizontally (e.g. each address receives every Cth 268 port spread across a pool of external IP addresses). 270 4: Cryptographically random port assignment (Section 2.2 of 271 RFC6431 [RFC6431]). If this algorithm is used, the Service 272 Provider needs to retain the keying material and specific 273 cryptographic function to support reversibility. 275 5: Vendor-specific. Other vendor-specific algorithms may also be 276 supported. 278 The assigned range of ports MAY also be used when translating ICMP 279 requests (when re-writing the Identifier field). 281 The CGN then reserves ports as follows: 283 1. The CGN removes reserved ports (R) from the port candidate list 284 (e.g., 0-1023 for TCP and UDP). At a minimum, the CGN SHOULD 285 remove system ports (RFC6335) [RFC6335] from the port candidate 286 list reserved for deterministic assignment. 288 2. The CGN calculates the total compression ratio (C+D), and 289 allocates 1/(C+D) of the available ports to each internal IP 290 address. Specific port allocation is determined by the algorithm 291 (A) configured on the CGN. Any remaining ports are allocated to 292 the dynamic pool. 294 Note: Setting D to 0 disables the dynamic pool. This option 295 eliminates the need for per-subscriber logging at the expense of 296 limiting the number of concurrent connections that 'power users' 297 can initiate. 299 3. When a subscriber initiates a connection, the CGN creates a 300 translation mapping between the subscriber's inside local IP 301 address/port and the CGN outside global IP address/port. The CGN 302 MUST use one of the ports allocated in step 2 for the translation 303 as long as such ports are available. The CGN SHOULD allocate 304 ports randomly within the port range assigned by the 305 deterministic algorithm. This is to increase subscriber privacy. 306 The CGN MUST use the preallocated port range from step 2 for Port 307 Control Protocol (PCP, [I-D.ietf-pcp-base]) reservations as long 308 as such ports are available. While the CGN maintains its mapping 309 table, it need not generate a log entry for translation mappings 310 created in this step. 312 4. If D>0, the CGN will have a pool of ports left for dynamic 313 assignment. If a subscriber uses more than the range of ports 314 allocated in step 2 (but fewer than the configured maximum ports 315 M), the CGN assigns a block of ports from the dynamic assignment 316 range for such a connection or for PCP reservations. The CGN 317 MUST log dynamically assigned port blocks to facilitate 318 subscriber-to-address mapping. The CGN SHOULD manage dynamic 319 ports as described in [I-D.tsou-behave-natx4-log-reduction]. 321 5. Configuration of reserved ports (e.g., system ports) is left to 322 operator configuration. 324 Thus, the CGN will maintain translation mapping information for all 325 connections within its internal translation tables; however, it only 326 needs to externally log translations for dynamically-assigned ports. 328 2.1. IPv4 Port Utilization Efficiency 330 For Service Providers requiring an aggressive address sharing ratio, 331 the use of the algorithmic mapping may impact the efficiency of the 332 address sharing. A dynamic port range allocation assignment is more 333 suitable in those cases. 335 2.2. Planning & Dimensioning 337 Unlike dynamic approaches, the use of the algorithmic mapping 338 requires more effort from operational teams to tweak the algorithm 339 (e.g., size of the port range, address sharing ratio, etc.). 340 Dedicated alarms SHOULD be configured when some port utilization 341 thresholds are fired so that the configuration can be refined. 343 The use of algorithmic mapping also affects geolocation. Changes to 344 the inside and outside address ranges (e.g. due to growth, address 345 allocation planning, etc.) would require external geolocation 346 providers to recalibrate their mappings. 348 2.3. Deterministic CGN Example 350 To illustrate the use of deterministic NAT, let's consider a simple 351 example. The operator configures an inside address range (I) of 352 100.64.0.0/28 [RFC6598] and outside address (O) of 203.0.113.1. The 353 dynamic address pool factor (D) is set to '2'. Thus, the total 354 compression ratio is 1:(14+2) = 1:16. Only the system ports (e.g. 355 ports < 1024) are reserved (R) . This configuration causes the CGN to 356 preallocate ((65536-1024)/16 =) 4032 TCP and 4032 UDP ports per 357 inside IPv4 address. For the purposes of this example, let's assume 358 that they are allocated sequentially, where 100.64.0.1 maps to 359 203.0.113.1 ports 1024-5055, 100.64.0.2 maps to 203.0.113.1 ports 360 5056-9087, etc. The dynamic port range thus contains ports 361 57472-65535 (port allocation illustrated in the table below). 362 Finally, the maximum ports/subscriber is set to 5040. 364 +-----------------------+-------------------------+ 365 | Inside Address / Pool | Outside Address & Port | 366 +-----------------------+-------------------------+ 367 | Reserved | 203.0.113.1:0-1023 | 368 | 100.64.0.1 | 203.0.113.1:1024-5055 | 369 | 100.64.0.2 | 203.0.113.1:5056-9087 | 370 | 100.64.0.3 | 203.0.113.1:9088-13119 | 371 | 100.64.0.4 | 203.0.113.1:13120-17151 | 372 | 100.64.0.5 | 203.0.113.1:17152-21183 | 373 | 100.64.0.6 | 203.0.113.1:21184-25215 | 374 | 100.64.0.7 | 203.0.113.1:25216-29247 | 375 | 100.64.0.8 | 203.0.113.1:29248-33279 | 376 | 100.64.0.9 | 203.0.113.1:33280-37311 | 377 | 100.64.0.10 | 203.0.113.1:37312-41343 | 378 | 100.64.0.11 | 203.0.113.1:41344-45375 | 379 | 100.64.0.12 | 203.0.113.1:45376-49407 | 380 | 100.64.0.13 | 203.0.113.1:49408-53439 | 381 | 100.64.0.14 | 203.0.113.1:53440-57471 | 382 | Dynamic | 203.0.113.1:57472-65535 | 383 +-----------------------+-------------------------+ 385 When subscriber 1 using 100.64.0.1 initiates a low volume of 386 connections (e.g. < 4032 concurrent connections), the CGN maps the 387 outgoing source address/port to the preallocated range. These 388 translation mappings are not logged. 390 Subscriber 2 concurrently uses more than the allocated 4032 ports 391 (e.g. for peer-to-peer, mapping, video streaming, or other 392 connection-intensive traffic types), the CGN allocates up to an 393 additional 1008 ports using bulk port reservations. In this example, 394 subscriber 2 uses outside ports 5056-9087, and then 100-port blocks 395 between 58000-58999. Connections using ports 5056-9087 are not 396 logged, while 10 log entries are created for ports 58000-58099, 397 58100-58199, 58200-58299, ..., 58900-58999. 399 In order to identify a subscriber behind a CGN (regardless of port 400 allocation method), public safety agencies need to collect source 401 address and port information from content provider log files. Thus, 402 content providers are advised to log source address, source port, and 403 timestamp for all log entries, per [RFC6302]. If a public safety 404 agency collects such information from a content provider and reports 405 abuse from 203.0.113.1, port 2001, the operator can reverse the 406 mapping algorithm to determine that the internal IP address 407 subscriber 1 has been assigned generated the traffic without 408 consulting CGN logs (by correlating the internal IP address with DHCP 409 /PPP lease connection records). If a second abuse report comes in 410 for 203.0.113.1, port 58204, the operator will determine that port 411 58204 is within the dynamic pool range, consult the log file, 412 correlate with connection records, and determine that subscriber 2 413 generated the traffic (assuming that the public safety timestamp 414 matches the operator timestamp. As noted in RFC6292 [RFC6292], 415 accurate time-keeping (e.g., use of NTP or Simple NTP) is vital). 417 In this example, there are no log entries for the majority of 418 subscribers, who only use pre-allocated ports. Only minimal logging 419 would be needed for those few subscribers who exceed their pre- 420 allocated ports and obtain extra bulk port assignments from the 421 dynamic pool. Logging data for those users will include inside 422 address, outside address, outside port range, and timestamp. 424 3. Additional Logging Considerations 426 In order to be able to identify a subscriber based on observed 427 external IPv4 address, port, and timestamp, an operator needs to know 428 how the CGN was configured with regards to internal and external IP 429 addresses, dynamic address pool factor, maximum ports per user, and 430 reserved port range at any given time. Therefore, the CGN MUST 431 generate a record any time such variables are changed. The CGN 432 SHOULD generate a log message any time such variables are changed. 433 The CGN MAY keep such a record in the form of a router configuration 434 file. If the CGN does not generate a log message, it would be up to 435 the operator to maintain version control of router config changes. 436 Also, the CGN SHOULD generate such a log message once per day to 437 facilitate quick identification of the relevant configuration in the 438 event of an abuse notification. 440 Such a log message MUST, at minimum, include the timestamp, inside 441 prefix I, inside mask, outside prefix O, outside mask, D, M, A, and 442 reserved port list R; for example: 444 [Wed Oct 11 14:32:52 445 2000]:100.64.0.0:28:203.0.113.0:32:2:5040:0:1-1023,5004,5060. 447 3.1. Failover Considerations 449 Due to the deterministic nature of algorithmically-assigned 450 translations, no additional logging is required during failover 451 conditions provided that inside address ranges are unique within a 452 given failover domain. Even when directed to a different CGN server, 453 translations within the deterministic port range on either the 454 primary or secondary server can be algorithmically reversed, provided 455 the algorithm is known. Thus, if 100.64.0.1 port 3456 maps to 456 203.0.113.1 port 1000 on CGN 1 and 198.51.100.1 port 1000 on Failover 457 CGN 2, an operator can identify the subscriber based on outside 458 source address and port information. 460 Similarly, assignments made from the dynamic overflow pool need to be 461 logged as described above, whether translations are performed on the 462 primary or failover CGN. 464 4. Impact on the IPv6 Transition 466 The solution described in this document is applicable to Carrier 467 Grade NAT transition technologies (e.g. NAT444, DS-Lite, and NAT64). 468 As discussed in [I-D.donley-nat444-impacts], the authors acknowledge 469 that native IPv6 will offer subscribers a better experience than CGN. 470 However, many CPE devices only support IPv4. Likewise, as of July 471 2012, only approximately 4% of the top 1 million websites were 472 available using IPv6. Accordingly, deterministic CGN should in no 473 way be understood as making CGN a replacement for IPv6 service. The 474 authors encourage device manufacturers to consider [RFC6540] and 475 include IPv6 support. In the interim, however, CGN has already been 476 deployed in some operator networks. Deterministic CGN will provide 477 operators with the ability to quickly respond to public safety 478 requests without requiring excessive infrastructure, operations, and 479 bandwidth to support per-connection logging. 481 5. IANA Considerations 483 This document makes no request of IANA. 485 6. Security Considerations 487 The security considerations applicable to NAT operation for various 488 protocols as documented in, for example, RFC 4787 [RFC4787] and RFC 489 5382 [RFC5382] also apply to this document. 491 Note that with the possible exception of cryptographically-based port 492 allocations, attackers could reverse-engineer algorithmically-derived 493 port allocations to either target a specific subscriber or to spoof 494 traffic to make it appear to have been generated by a specific 495 subscriber. However, this is exactly the same level of security that 496 the subscriber would experience in the absence of CGN. CGN is not 497 intended to provide additional security by obscurity. 499 7. Acknowledgements 501 The authors would like to thank the following people for their 502 suggestions and feedback: Bobby Flaim, Lee Howard, Wes George, Jean- 503 Francois Tremblay, Mohammed Boucadair, Alain Durand, David Miles, 504 Andy Anchev, Victor Kuarsingh, Miguel Cros Cecilia, and Reinaldo 505 Penno. 507 8. References 509 8.1. Normative References 511 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 512 Requirement Levels", BCP 14, RFC 2119, March 1997. 514 [RFC4787] Audet, F. and C. Jennings, "Network Address Translation 515 (NAT) Behavioral Requirements for Unicast UDP", BCP 127, 516 RFC 4787, January 2007. 518 [RFC5382] Guha, S., Biswas, K., Ford, B., Sivakumar, S., and P. 519 Srisuresh, "NAT Behavioral Requirements for TCP", BCP 142, 520 RFC 5382, October 2008. 522 [RFC6264] Jiang, S., Guo, D., and B. Carpenter, "An Incremental 523 Carrier-Grade NAT (CGN) for IPv6 Transition", RFC 6264, 524 June 2011. 526 [RFC6269] Ford, M., Boucadair, M., Durand, A., Levis, P., and P. 527 Roberts, "Issues with IP Address Sharing", RFC 6269, June 528 2011. 530 8.2. Informative References 532 [I-D.donley-nat444-impacts] 533 Donley, C., Howard, L., Kuarsingh, V., Berg, J., and U. 534 Colorado, "Assessing the Impact of Carrier-Grade NAT on 535 Network Applications", draft-donley-nat444-impacts-04 536 (work in progress), May 2012. 538 [I-D.ietf-pcp-base] 539 Wing, D., Cheshire, S., Boucadair, M., Penno, R., and P. 540 Selkirk, "Port Control Protocol (PCP)", draft-ietf-pcp- 541 base-29 (work in progress), November 2012. 543 [I-D.miles-behave-l2nat] 544 Miles, D. and M. Townsley, "Layer2-Aware NAT", draft- 545 miles-behave-l2nat-00 (work in progress), March 2009. 547 [I-D.shirasaki-nat444] 548 Yamagata, I., Shirasaki, Y., Nakagawa, A., Yamaguchi, J., 549 and H. Ashida, "NAT444", draft-shirasaki-nat444-06 (work 550 in progress), July 2012. 552 [I-D.sivakumar-behave-nat-logging] 553 Sivakumar, S. and R. Penno, "IPFIX Information Elements 554 for logging NAT Events", draft-sivakumar-behave-nat- 555 logging-05 (work in progress), July 2012. 557 [I-D.tsou-behave-natx4-log-reduction] 558 ZOU), T., Li, W., and T. Taylor, "Port Management To 559 Reduce Logging In Large-Scale NATs", draft-tsou-behave- 560 natx4-log-reduction-02 (work in progress), September 2010. 562 [RFC6056] Larsen, M. and F. Gont, "Recommendations for Transport- 563 Protocol Port Randomization", BCP 156, RFC 6056, January 564 2011. 566 [RFC6146] Bagnulo, M., Matthews, P., and I. van Beijnum, "Stateful 567 NAT64: Network Address and Protocol Translation from IPv6 568 Clients to IPv4 Servers", RFC 6146, April 2011. 570 [RFC6292] Hoffman, P., "Requirements for a Working Group Charter 571 Tool", RFC 6292, June 2011. 573 [RFC6302] Durand, A., Gashinsky, I., Lee, D., and S. Sheppard, 574 "Logging Recommendations for Internet-Facing Servers", BCP 575 162, RFC 6302, June 2011. 577 [RFC6333] Durand, A., Droms, R., Woodyatt, J., and Y. Lee, "Dual- 578 Stack Lite Broadband Deployments Following IPv4 579 Exhaustion", RFC 6333, August 2011. 581 [RFC6335] Cotton, M., Eggert, L., Touch, J., Westerlund, M., and S. 582 Cheshire, "Internet Assigned Numbers Authority (IANA) 583 Procedures for the Management of the Service Name and 584 Transport Protocol Port Number Registry", BCP 165, RFC 585 6335, August 2011. 587 [RFC6431] Boucadair, M., Levis, P., Bajko, G., Savolainen, T., and 588 T. Tsou, "Huawei Port Range Configuration Options for PPP 589 IP Control Protocol (IPCP)", RFC 6431, November 2011. 591 [RFC6540] George, W., Donley, C., Liljenstolpe, C., and L. Howard, 592 "IPv6 Support Required for All IP-Capable Nodes", BCP 177, 593 RFC 6540, April 2012. 595 [RFC6598] Weil, J., Kuarsingh, V., Donley, C., Liljenstolpe, C., and 596 M. Azinger, "IANA-Reserved IPv4 Prefix for Shared Address 597 Space", BCP 153, RFC 6598, April 2012. 599 Authors' Addresses 601 Chris Donley 602 CableLabs 603 858 Coal Creek Cir 604 Louisville, CO 80027 605 US 607 Email: c.donley@cablelabs.com 609 Chris Grundemann 610 Internet Society 611 Denver, CO 612 US 614 Email: cgrundemann@gmail.com 616 Vikas Sarawat 617 CableLabs 618 858 Coal Creek Cir 619 Louisville, CO 80027 620 US 622 Email: v.sarawat@cablelabs.com 624 Karthik Sundaresan 625 CableLabs 626 858 Coal Creek Cir 627 Louisville, CO 80027 628 US 630 Email: k.sundaresan@cablelabs.com 632 Olivier Vautrin 633 Juniper Networks 634 1194 N Mathilda Avenue 635 Sunnyvale, CA 94089 636 US 638 Email: olivier@juniper.net