idnits 2.17.1 draft-ietf-opsawg-lsn-deployment-00.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 : ---------------------------------------------------------------------------- ** The abstract seems to contain references ([RFC4364], [I-D.ietf-behave-lsn-requirements]), which it shouldn't. Please replace those with straight textual mentions of the documents in question. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (May 15, 2012) is 4365 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- == Outdated reference: A later version (-10) exists of draft-ietf-behave-lsn-requirements-06 Summary: 1 error (**), 0 flaws (~~), 2 warnings (==), 1 comment (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 OPSAWG V. Kuarsingh, Ed. 3 Internet-Draft J. Cianfarani 4 Intended status: Informational Rogers Communications 5 Expires: November 16, 2012 May 15, 2012 7 CGN Deployment with BGP/MPLS IP VPNs 8 draft-ietf-opsawg-lsn-deployment-00 10 Abstract 12 This document specifies a framework to integrate a Network Address 13 Translation layer into an operator's network to function as a Carrier 14 Grade NAT (also known as CGN or Large Scale NAT). CGN is a concept 15 also described in [I-D.ietf-behave-lsn-requirements] and describes 16 the model as a dual layer translation model. Although operators may 17 wish to deploy IPv6 to strategically overcome IPv4 exhaustion, near 18 term needs may not be satisfied with an IPv6 deployment alone. This 19 document provides a practical integration model which allows CGN to 20 be integrated into the network meeting the connectivity needs of the 21 customer while being mindful of not disrupting existing services and 22 meeting the technical challenges that CGN brings. The model includes 23 the use of BGP/MPLS IP VPNs defined in [RFC4364] as a tool to achieve 24 this goal. This document does not intend to defend the merits of 25 CGN. 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 http://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 November 16, 2012. 44 Copyright Notice 46 Copyright (c) 2012 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 51 (http://trustee.ietf.org/license-info) in effect on the date of 52 publication of this document. Please review these documents 53 carefully, as they describe your rights and restrictions with respect 54 to this document. Code Components extracted from this document must 55 include Simplified BSD License text as described in Section 4.e of 56 the Trust Legal Provisions and are provided without warranty as 57 described in the Simplified BSD License. 59 Table of Contents 61 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 3 62 2. Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . 3 63 3. CGN Network Deployment Requirements . . . . . . . . . . . . . 4 64 3.1. Centralized versus Distributed Deployment . . . . . . . . 5 65 3.2. CGN and Traditional IPv4 Service Co-existence . . . . . . 6 66 3.3. CGN By-Pass . . . . . . . . . . . . . . . . . . . . . . . 6 67 3.4. Routing Plane Separation . . . . . . . . . . . . . . . . . 6 68 3.5. Flexible Deployment Options . . . . . . . . . . . . . . . 7 69 3.6. IPv4 Overlap Space . . . . . . . . . . . . . . . . . . . . 7 70 3.7. Transactional Logging for LSN Systems . . . . . . . . . . 7 71 3.8. Additional CGN Requirements . . . . . . . . . . . . . . . 8 72 4. BGP/MPLS IP VPN based CGN Framework . . . . . . . . . . . . . 8 73 4.1. Service Separation . . . . . . . . . . . . . . . . . . . . 9 74 4.2. Internal Service Delivery . . . . . . . . . . . . . . . . 10 75 4.2.1. Dual Stack Operation . . . . . . . . . . . . . . . . . 11 76 4.3. Deployment Flexibility . . . . . . . . . . . . . . . . . . 13 77 4.4. Comparison of BGP/MPLS IP VPN Option versus other CGN 78 Attachment Options . . . . . . . . . . . . . . . . . . . . 13 79 4.4.1. IEEE 802.1Q . . . . . . . . . . . . . . . . . . . . . 13 80 4.4.2. Policy Based Routing . . . . . . . . . . . . . . . . . 14 81 4.4.3. Traffic Engineering . . . . . . . . . . . . . . . . . 14 82 4.4.4. Multiple Routing Topologies . . . . . . . . . . . . . 14 83 5. Experiences . . . . . . . . . . . . . . . . . . . . . . . . . 14 84 6. Basic Integration and Requirements Support . . . . . . . . . . 14 85 7. Performance . . . . . . . . . . . . . . . . . . . . . . . . . 15 86 8. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 16 87 9. Security Considerations . . . . . . . . . . . . . . . . . . . 16 88 10. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . 17 89 11. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . 17 90 12. References . . . . . . . . . . . . . . . . . . . . . . . . . . 17 91 12.1. Normative References . . . . . . . . . . . . . . . . . . . 17 92 12.2. Informative References . . . . . . . . . . . . . . . . . . 17 93 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . . 18 95 1. Introduction 97 Operators are faced with near term IPv4 address exhaustion 98 challenges. Many operators may not have a sufficient amount of IPv4 99 addresses in the future to satisfy the needs of their growing 100 customer base. This challenge may also be present before or during 101 an active transition to IPv6 somewhat complicating the overall 102 problem space. 104 To face this challenge, operators may need to deploy CGN (Carrier 105 Grade NAT) as described in [I-D.ietf-behave-lsn-requirements] to help 106 extend the connectivity matrix once IPv4 addresses run out in the 107 network. CGN's addition to the network requires integration in an 108 often running state environment with working IPv4 and/or IPv6 109 services. 111 The addition of the CGN introduces an operator controlled and 112 administered translation layer which needs to be added in a manner 113 which does not overly disrupt existing services. This addition may 114 also include interworking in a dual stack environment where the IPv4 115 path requires translation. 117 This document shows how BGP/MPLS IP VPNs as described in [RFC4364] 118 can be used to integrate the CGN infrastructure solving key problems 119 faced by the operator. This model has also been tested and validated 120 in real production network models and allows fluid operation with 121 existing IPv4 and IPv6 services. 123 2. Motivation 125 The selection of CGN may be made by an operator based on a number of 126 factors. The overall driver may be the depletion of IPv4 address 127 pools which leaves little to no addresses for IPv4 service growth. 128 IPv6 is considered the strategic answer, but it's applicability and 129 usefulness in many networks is limited by the current access network 130 and consumer home network. These environments often are filled with 131 IPv4-Only equipment which may not be upgradable to IPv6. 133 The ability to replace IPv4-Only equipment may be out of the control 134 of the operator, and even when it's in the administrative control; it 135 poses both cost and technical challenges as operators build out 136 massive programs for equipment retirement or upgrade. Theses issues 137 leave an operator in a precarious position which may lead to the 138 decision to deploy CGN. Other address IPv4 sharing options do exist 139 which are more architecturally desirable, but the practical and 140 workable approach in many cases is a CGN deployment using NAT444. 142 If the operator as has chosen to deploy CGN, they should this in a 143 manner as not to negatively impact the existing IPv4 or IPv6 customer 144 base. This will include solving a number of challenges since 145 customers who's connections require translation will have network 146 routing and flow needs which are different from legacy IPv4 147 connections. 149 The solution will also need to work in a dual stack environment where 150 other options such as DS-Lite [RFC6333] are not yet viable. Even 151 technologies like 6RD [RFC5969] still require an IPv4 connectivity 152 path to service the customer endpoint. The solution will need to 153 address basic Internet connectivity, on-net service offerings, back 154 office management, billing, policy and security models already in 155 place within the operator's network. CGN will often integrate quite 156 readily with the aforementioned requirements where as other 157 transition mechanism may not due to the requirements to support IPv6 158 as the base protocol for IPv4 connectivity. 160 3. CGN Network Deployment Requirements 162 If a service provider is considering a CGN deployment with a provider 163 NAT44 function, there are a number of basic requirements which are of 164 importance. Preliminary requirements may require the following from 165 the incoming CGN system architecture: 167 - Support distributed (sparse) and centralized (dense) deployment 168 models; 170 - Allow co-existence with traditional IPv4 based deployments, 171 which provide global scoped IPs to CPEs; 173 - Provide a framework for CGN by-pass supporting non-translated 174 flows between endpoints within a provider's network; 176 - Provide routing framework which allows the segmentation of 177 routing control and forwarding paths between CGN and non-CGN 178 mediated flows; 180 - Provide flexibility for operators to modify their deployments 181 over time as translation demands change (connections, bandwidth, 182 translation realms/zones and other vectors); 184 - Flexibility should include integration options for common access 185 technologies such as DSL (BRAS), DOCSIS (CMTS), Mobile (GGSN/PGW/ 186 ASN-GW), and Ethernet access; 187 - Support deployment modes that allow for IPv4 address overlap 188 within the operator's network (between various translation realms 189 or zones); 191 - Allow for evolution to future dual-stack and IPv4/IPv6 192 transition deployment modes; 194 - Transactional logging and export capabilities to support 195 auxiliary functions including abuse mitigation; 197 - Support for stateful connection synchronization between 198 translation instances/elements (redundancy); 200 - Support for CGN Shared Space [RFC6598] deployment modes if 201 applicable; 203 - Allows for the enablement of CGN functionality (if required) 204 while still minimizing costs and customer impact to the best 205 extend possible; 207 Other requirements may be assessed on a operator-by-operator basis, 208 but those listed above should be considered for any given deployment 209 architecture. 211 3.1. Centralized versus Distributed Deployment 213 Centralized deployments of CGN (longer proximity to end user and/or 214 higher densities of subscribers/connections to CGN instances) differ 215 from distributed deployments of CGN (closer proximity to end user 216 and/or lower densities of subscribers/connections to CGN instances). 217 Service providers will likely deploy CGN translation points more 218 centrally during initial phases. Early deployments will likely see 219 light loading on these new systems since legacy IPv4 services will 220 continue to operate with most endpoints using globally unique IPv4 221 addresses. Exceptional cases which may drive heavy usage in initial 222 stages may include operators who already translate most IPv4 traffic 223 and will migrate to a CGN implementation from legacy firewalls; or a 224 green field deployment which may see quick growth in the number of 225 new IPv4 endpoints which require Internet connectivity. 227 Over time, most providers will likely need to expand and possibly 228 distribute the translation points as demand for the CGN system 229 increases. The extent of the expansion of the CGN infrastructure 230 will depend on factors such as growth in the number of IPv4 231 endpoints, status of IPv6 content on the Internet and the overall 232 progress globally to an IPv6-dominate Internet (reducing the demand 233 for IPv4 connectivity). 235 3.2. CGN and Traditional IPv4 Service Co-existence 237 Newer CGN serviced endpoints will exist alongside endpoints served by 238 traditional IPv4 global IPs. Providers will need to rationalize 239 these environments since both have distinct forwarding needs. 240 Traditional IPv4 services will likely require (or be best served) 241 direct forwarding towards Internet peering points while CGN mediated 242 flows require access to a translator. CGN and non-CGN mediated flows 243 post two fundamentally different forwarding needs. 245 The new CGN environments should not negatively impact the existing 246 IPv4 service base by forcing all traffic to translation enabled 247 network points since many flows do not require translation and this 248 would reduce performance of the existing flows. This would also 249 require massive scaling of the CGN which is a cost and efficiency 250 concern as well. 252 Traffic flow and forwarding efficiency is considered important since 253 networks are under considerable demand to deliver more and more 254 bandwidth without the luxury of needless inefficiencies which can be 255 introduced with CGN. 257 3.3. CGN By-Pass 259 The CGN environment is only needed for flows with translation 260 requirements. Many flows which remain in a service provider 261 environment, do not require translation. Such services include 262 operator offered DNS Services, DHCP Services, NTP Services, Web 263 Caching, Mail, News and other services which are local to the 264 operator's network. 266 The operator may want to leverage opportunities to offer third 267 parties a platform to also provide services without translation. CGN 268 By-pass can be accomplished in many ways, but a simplistic, 269 deterministic and scalable model is preferred. 271 3.4. Routing Plane Separation 273 Many operators will want to engineer traffic separately for CGN flows 274 versus flows which are part of the more traditional IPv4 environment. 275 Many times the routing of these two major flow types differ, 276 therefore route separation may be required. 278 Routing plane separation also allows the operator to utilize other 279 addressing techniques, which may not be feasible on a single routing 280 plane. Such examples include the use of overlapping private address 281 space [RFC1918] or use of other IPv4 space which may overlap globally 282 within the operator's network. 284 3.5. Flexible Deployment Options 286 Service providers operate complex routing environments and offer a 287 variety of IPv4 based services. Many operator environments utilize 288 distributed peering infrastructures for transit and peering and these 289 may span large geographical areas and regions. A CGN solution should 290 offer the operator an ability to place CGN translation points at 291 various points within their network. 293 The CGN deployment should also be flexible enough to change over time 294 as demand for translation services increase. In turn, the deployment 295 will need to then adapt as translation demand decreases caused by the 296 transition of flows to IPv6. Translation points should be able to be 297 placed and moved with as little re-engineering effort as possible 298 minimizing the risks to the customer base. 300 Depending on hardware capabilities, security practices and IPv4 301 address availability, the translation environments my need to be 302 segmented and/or scaled over time to meet organic IPv4 demand growth. 303 Operators will want to seek deployment models which are conducive to 304 meeting these goals as well. 306 3.6. IPv4 Overlap Space 308 IP address overlap for CGN translation realms may be required if 309 insufficient IPv4 addresses are available within the service provider 310 environment to assign internally unique IPs to the CGN customer base 311 . The CGN deployment should provide mechanisms to manage IPv4 312 overlap if required. 314 3.7. Transactional Logging for LSN Systems 316 CGNs may require transactional logging since the source IP and 317 related transport protocol information is not easily visible to 318 external hosts and system. 320 If needed, the CGN systems should be able to generate logs which 321 identify 'internal' host parameters (i.e. IP/Port) and associated 322 them to external translated parameters imposed by the translator. 323 The logged information should be stored on the CGN hardware and/or 324 exported to an external system for processing. Operators may need to 325 keep track of this information (securely) to meet regulatory and/or 326 legal obligations. Further information can be found in [I-D.ietf- 327 behave-lsn-requirements] with respect to CGN logging requirements 328 (Logging Section). 330 3.8. Additional CGN Requirements 332 The CGN platform will also need to meet the needs of additional 333 requirements such as Bulk Port Allocation and other CGN device 334 specific functions. These additional requirements are captured 335 within [I-D.ietf-behave-lsn-requirements]. 337 4. BGP/MPLS IP VPN based CGN Framework 339 The BGP/MPLS IP VPN [RFC4364] framework for CGN segregates the 'pre- 340 translated' realms within the service provider space into Layer-3 341 MPLS based VPNs. The operator can deploy a single realm for all CGN 342 based flows, or can deploy multiple realms based on translation 343 demand and other factors such as geographical proximity. A realm in 344 this model refers to a 'VPN' which shares a unique RD/RT combination, 345 routing plane and forwarding behaviours. 347 The BGP/MPLS IP VPN infrastructure provides control plane and 348 forwarding separation for the traditional IPv4 service environment 349 and CGN environment(s). The separation allows for routing 350 information (such as default routes) to be propagated separately for 351 CGN and non-CGN based customer flows. Traffic can be efficiently 352 routed to the Internet for normal flows, and routed directly to 353 translators for CGN mediated flows. Although many operators may run 354 a "default-route-free" core, IPv4 flows which require translation 355 must obviously be routed first to a translator, so a default route is 356 acceptable for the pre-translated realms. 358 The physical location of the VRF Termination point for a BGP/MPLS IP 359 VPN enabled CGN can vary and be located anywhere within the 360 operator's network. This model fully virtualizes the translation 361 service from the base IPv4 forwarding environment which will likely 362 carrying Internet bound traffic. The base IPv4 environment can 363 continue to service traditional IPv4 customer flows plus post 364 translated CGN flows. 366 Figure 1 provides a view of the basic model. The Access node 367 provides CPE access to either the CGN VRF or the Global Routing 368 Table, depending on whether the customer receives a private or public 369 IP. Translator mediated traffic follows an MPLS LSP which can be 370 setup dynamically and can span one hop, or many hops (with no need 371 for complex routing policies). Traffic is then forwarded to the 372 translator (shown below) which can be an external appliance or 373 integrated into the VRF Termination (Provider Edge) router. Once 374 traffic is translated, it is forwarded to the global routing table 375 for general Internet forwarding. The Global Routing table can also 376 be a separate VRF (Internet Access VPN/VRF) should the provider 377 choose to implement their Internet based services in that fashion. 378 The translation services are effectively overlaid onto the network, 379 but are maintained within a separate forwarding and control plane. 381 Access Node VRF Termination LSN 382 +-----------+ +-----------+ +-----------+ 383 | | | | | | 384 CPE | +-------+ | | +-------+ | | +-------+ | 385 +----+ | | | | LSP | | | | IP | | | | 386 | --+---+-+->VRF--+-+-----+-+->VRF--+-+----+-+-> | | 387 +----+ | | | | | | | | | | | | 388 | +-------+ | | +-------+ | | | | | 389 | | | | | | XLATE | | 390 | | | | | | | | 391 CPE | +-------+ | | +-------+ | | | | | 392 +----+ | | | | | | | | IP | | | | 393 | --+---+-+->GRT | | | | GRT<-+-+----+-+-- | | 394 +----+ | | | | | | | | | | | | | | 395 | +---+---+ | | +---+---+ | | +-------+ | 396 +-----+-----+ +-----+-----+ +-----------+ 397 | | 398 | | IPv4 399 | | IP +---------+ 400 | +------------+-> | 401 | IP | GRT | 402 +------------------------------+-> | 403 +---------+ 405 Figure 1: Basic BGP/MPLS IP VPN CGN Model 407 If more then one VRF (translation realm) is used within the 408 operator's network, each VPN instance can manage CGN flows 409 independently for the respective realm. Various redundancy models 410 can be used within this architecture to support failover from one 411 physical CGN hardware instance to another. If state information 412 needs to be passed or maintained between hardware instances, the 413 vendor would need to enable this feature in a suitable manner. 415 4.1. Service Separation 417 The MPLS/VPN CGN framework supports route separation. The 418 traditional IPv4 flows can be separated at the access node (Initial 419 Layer 3 service point) from those which require translation. This 420 type of service separation is possible on common technologies used 421 for Internet access within many operator networks. Service 422 separation can be accomplished on common access technology including 423 those used for DOCSIS (CMTS), Ethernet Access, DSL (BRAS), and Mobile 424 Access (GGSN/ASN-GW) architectures. 426 4.2. Internal Service Delivery 428 Internal services can be delivered directly to the privately 429 addressed endpoint within the CGN domain without translation. This 430 can be accomplished using direct route exchange (import/export) 431 between the CGN VRFs and the Services VRFs. The previous statement 432 assumes the provider puts key services into a VRF for simple route 433 exchange. This model allows the provider to maintain separate 434 forwarding rules for translated flows, which require a pass through 435 the translator to reach external network entities, versus those flows 436 which need to access internal services. This operational detail can 437 be advantageous for a number of reasons. 439 First, the provider can reduce the load on the translator since 440 internal services do not need to be factored into the scaling of the 441 CGN hardware. Secondly, more direct forwarding paths can be 442 maintained providing better network efficiency. Thirdly, geographic 443 locations of the translators and the services infrastructure can be 444 deployed in a location in an independent manner. Additionally, the 445 operator can allow CGN subject endpoints to be accessible via an 446 untranslated path reducing the complexities of provider initiated 447 management flows. This last point is of key interest since NAT 448 removes transparency to the end device in normal cases. 450 Figure 2 below shows how internal services are provided untranslated 451 since flows are sent directly from the access node to the services 452 node/VRF via an MPLS LSP. This traffic is not forwarded to the CGN 453 translator and therefore is not subject to problematic behaviours 454 related to NAT. The services VRF contains routing information which 455 can be "imported" into the access node VRF and the CGN VRF routing 456 information can be "imported" into the Services VRF. 458 Access Node VRF Termination LSN 459 +-------------+ +-----------+ +----------+ 460 | | | | | | 461 CPE | +---------+ | | +-------+ | | +------+ | 462 +-----+ | | | | | | | | | | | | 463 | --+--+-+-> VRF --+-+--+ | | VRF | | | | | | 464 +-----+ | | | | | | | | | | | | | 465 | +---------+ | | | +-------+ | | | | | 466 | | | | | | |XLATE | | 467 | | | | | | | | | 468 CPE | +---------+ | | | +-------+ | | | | | 469 +-----+ | | | | | | | | | | | | | 470 | --+--+-+-> GRT | | | | | GRT | | | | | | 471 +-----+ | | | | | | | | | | | | | | 472 | +----+----+ | | | +-------+ | | +------+ | 473 +------+------+ | +-----------+ +----------+ 474 | | 475 | | IPv4 476 | | +-----------+ 477 | +---------------+->Services | 478 | | VRF | 479 .-------------------------+-> | | 480 +-----+-----+ 481 | 482 +-----V-----+ 483 | | 484 | Local | 485 | Content | 486 +-----------+ 488 Figure 2: Internal Services and CGN By-Pass 490 This demonstrates the ability to offer CGN By-Pass in a simple and 491 deterministic manner without the need of policy based routing or 492 traffic engineering. 494 4.2.1. Dual Stack Operation 496 The BGP/MPLS IP VPN CGN model can also be used in conjunction with 497 IPv4/IPv6 dual stack service modes. Since many providers will use 498 CGNs on an interim basis while IPv6 matures within the global 499 Internet or due to technical constraints, a dual stack option is of 500 strategic importance. Operators can offer this dual stack service 501 for both traditional IPv4 (global IP) endpoints and CGN mediated 502 endpoints. 504 Operators can separate the IP flows for IPv4 and IPv6 traffic, or use 505 other routing techniques to move IPv6 based flows towards the GRT 506 (Global Routing Table or Instance) while allowing IPv4 flows to 507 remain within the IPv4 CGN VRF for translator services. 509 The Figure 3 below shows how IPv4 translation services can be 510 provided alongside IPv6 based services. The model shown allows the 511 provider to enable CGN to manage IPv4 flows (translated) and IPv6 512 flows are routed without translation efficiently towards the 513 Internet. Once again, forwarding of flows to the translator does not 514 impact IPv6 flows which do not require this service. 516 Access Node VRF Termination LSN 517 +-----------+ +-----------+ +-----------+ 518 | | | | | | 519 CPE | +-------+ | | +-------+ | | +-------+ | 520 +-----+ | | | |LSP| | | | IP | | | | 521 | --+--+-+->VRF--+-+---+-+->VRF--+-+----+-+> | | 522 |IPv4 | | | | | | | | | | | | | 523 | | | +-------+ | | +-------+ | | | | | 524 +-----| | | | | | | XLATE | | 525 |IPv6 | | | | | | | | | 526 | | | +-------+ | | +-------+ | | | | | 527 | | | | IPv6 | | | | IPv4 | | IP | | | | 528 | --+--+-+->GRT | | | | GRT<-+-+----+-+-- | | 529 +-----+ | | | | | | | | | | | | | | 530 | +---+---+ | | +---+---+ | | +-------+ | 531 +-----+-----+ +-----+-----+ +-----------+ 532 | | 533 | | +-----------+ 534 | | IP | IPv4 | 535 | +----------+-> GRT | 536 | +-----------+ 537 | 538 | 539 | 540 | IP +-----------+ 541 +--------------------------+-> IPv6 | 542 | GRT | 543 +-----------+ 545 Figure 3: CGN with IPv6 Dual Stack Operation 547 4.3. Deployment Flexibility 549 The CGN translator services can be moved, separated or segmented (new 550 translation realms) without the need to change the overall 551 translation design. Since dynamic LSPs are used to forward traffic 552 from the access nodes to the translation points, the physical 553 location of the VRF termination points can vary and be changed 554 easily. 556 This type of flexibility allows the service provider to initially 557 deploy more centralized translation services based on relatively low 558 loading factors, and distribute the translation points over time to 559 improve network traffic efficiencies and support higher translation 560 load. 562 Although traffic engineered paths are not required within the MPLS/ 563 VPN deployment model, nothing precludes an operator from using 564 technologies like MPLS with Traffic Engineering [RFC3031]. 565 Additional routing mechanisms can be used as desired by the provider 566 and can be seen as independent. There is no specific need to 567 diversify the existing infrastructure in most cases. 569 4.4. Comparison of BGP/MPLS IP VPN Option versus other CGN Attachment 570 Options 572 Other integration architecture options exist which can attach CGN 573 based service flows to a translator instance. Alternate options 574 which can be used to attach such services include: 576 - IEEE 802.1Q for direct attachment to a next hop translator; 578 - Policy Based Routing (Static) to direct translation bound 579 traffic to a network based translator; 581 - Traffic Engineering or; 583 - Multiple Routing Topologies 585 4.4.1. IEEE 802.1Q 587 IEEE 802.1Q can be used to associate separated traffic from the 588 access node to the next hop router's CGN instance. This technology 589 option may limit the CGN placement to the next hop router unless a 590 second technology option is paired with it to extend connectivity 591 deeper in the network. 593 This option is most effective if CGN instances are placed directly 594 upstream of the access node. Distributed CGN instance placement is 595 not likely an initial stage of the CGN deployment due to cost and 596 demand factors. 598 4.4.2. Policy Based Routing 600 Policy Based Routing (PBR) provides another option to direct CGN 601 mediated flows to a translator. PBR options, although possible, are 602 difficult to maintain (static policy) and must be configured 603 throughout the network with considerable maintenance overhead. 605 More centralized deployments may be difficult or too onerous to 606 deploy using Policy Based Routing methods. Policy Based Routing 607 would not achieve route separation (unless used with others options), 608 and may add complexities to the providers' routing environment. 610 4.4.3. Traffic Engineering 612 Traffic Engineering can also be used to direct traffic from an access 613 node towards a translator. Traffic Engineering, like MPLS-TE, may be 614 difficult to setup and maintain. Traffic Engineering provides 615 additional benefits if used with MPLS by adding potentials for faster 616 path re-convergence. Traffic Engineering paths would need to be 617 updated and redefined overtime as CGN translation points are 618 augmented or moved. 620 4.4.4. Multiple Routing Topologies 622 Multiple routing topologies can be used to direct CGN based flows to 623 translators. This option would achieve the same basic goal as the 624 MPLS/VPN option but with additional implementation overhead and 625 platform configuration complexity. Since operator based translation 626 is expected to have an unknown lifecycle, and may see various degrees 627 of demand (dependant on operator IPv4 Global space availability and 628 shift of traffic to IPv6), it may be too large of an undertaking for 629 the provider to enabled this as their primary option for CGN. 631 5. Experiences 633 6. Basic Integration and Requirements Support 635 The MPLS/VPN CGN environment has been successfully integrated into 636 real network environments utilizing existing network service delivery 637 mechanisms. It solves many issues related to provider based 638 translation environments, while still subject to problematic 639 behaviours inherent within NAT. 641 Key issues which are solved or managed with the MPLS/VPN option 642 include: 644 - Centralized and Distributed Deployment model support 646 - Routing Plane Separation for CGN flows versus traditional IPv4 647 flows 649 - Flexible Translation Point Design (can relocate translators and 650 split translation zones easily) 652 - Low maintenance overhead (dynamic routing environment with 653 little maintenance of separate routing infrastructure other then 654 management of MPLS/VPNs) 656 - CGN By-pass options (for internal and third party services which 657 exist within the provider domain) 659 - IPv4 Translation Realm overlap support (can reuse IP addresses 660 between zones with some impact to extranet service model) 662 - Simple failover techniques can be implemented with redundant 663 translators, such as using a second default route 665 7. Performance 667 The MPLS/VPN CGN model was observed to support basic functions which 668 are typically used by customers within an operator environment. 669 Examples of successful operation include: 671 - Traditional Web (HTTP) Surfing (client initiated) 673 - Internet Video Streaming 675 - HTTP Based Client Connections 677 - High Connection Count sites (i.e. Google Maps) 679 - Email Transaction Support (POP, IMAP, SMTP) 681 - Instant Messaging Support (Online Status, File transfers, text 682 chat) 684 - ICMP Operation (client initiated Echo, Traceroute) 686 - Peer to Peer application support (download) 687 - DNS (based on services extranet option, but was problematic when 688 passed through a translator) 690 CGNs are still subject to problematic connectivity even within the 691 MPLS/VPN technology approach. Problems which arise, or are not 692 inherently addressed in this model include: 694 - Inward services from the Internet to the CPE 696 - Web session tracking 698 - Restricting usage and/or access based on source IP 700 - Abuse mitigation (masquerade of potential offenders) 702 - Increased network or server IDS false positives 704 - Increased customer risk for session hijacking 706 - Exceeding firewall TCP/UDP limits 708 - Customer identification (external site) 710 - Poor source based load balancing 712 - Customer usage tracking / Ad insertion 714 - Other applications or operations may be negatively impacted 716 8. IANA Considerations 718 There are not specific IANA considerations known at this time with 719 the architecture described herein. Should a provide choose to use 720 non-assigned IP address space within their translation realms, then 721 considerations may apply. 723 9. Security Considerations 725 The same security considerations would typically exist for CGN 726 deployments when compared with traditional IPv4 based services. With 727 the MPLS/VPN model, the operator would want to consider security 728 issues related to offering IP services over MPLS. 730 If a provider plans to operate the pre-translation realm (CPE towards 731 translator IPv4 zone) as a non-public like network, then additional 732 security measures may be needed to secure this environment. It is 733 however the position in this document that CGN realms are public 734 domains which utilize non-Internet routable IP addresses for endpoint 735 addressing. 737 10. Conclusions 739 The MPLS/VPN delivery method for a CGN deployment is an effective and 740 scalable way to deliver mass translation services. The architecture 741 avoids the complex requirements of traffic engineering and policy 742 based routing when combining these new service flows to existing IPv4 743 operation. This is advantageous since the NAT44/CGN environments 744 should be introduced with as little impact as possible and these 745 environments are expected to change over time. 747 The MPLS/VPN based CGN architecture solves many of this issues 748 related to deploying this technology in existing operator networks. 750 11. Acknowledgements 752 Thanks to the following people for their participating in integrating 753 and testing the CGN environment: Chris Metz, Syd Alam, Richard 754 Lawson, John E Spence. 756 Additional thanks for the following people for the guidance on IPv6 757 transition considerations: John Jason Brzozowski, Chris Donley, Jason 758 Weil, Lee Howard, Jean-Francois Tremblay 760 12. References 762 12.1. Normative References 764 [RFC4364] Rosen, E. and Y. Rekhter, "BGP/MPLS IP Virtual Private 765 Networks (VPNs)", RFC 4364, February 2006. 767 12.2. Informative References 769 [I-D.ietf-behave-lsn-requirements] 770 Perreault, S., Yamagata, I., Miyakawa, S., Nakagawa, A., 771 and H. Ashida, "Common requirements for Carrier Grade NATs 772 (CGNs)", draft-ietf-behave-lsn-requirements-06 (work in 773 progress), May 2012. 775 [RFC1918] Rekhter, Y., Moskowitz, R., Karrenberg, D., Groot, G., and 776 E. Lear, "Address Allocation for Private Internets", 777 BCP 5, RFC 1918, February 1996. 779 [RFC3031] Rosen, E., Viswanathan, A., and R. Callon, "Multiprotocol 780 Label Switching Architecture", RFC 3031, January 2001. 782 [RFC5969] Townsley, W. and O. Troan, "IPv6 Rapid Deployment on IPv4 783 Infrastructures (6rd) -- Protocol Specification", 784 RFC 5969, August 2010. 786 [RFC6333] Durand, A., Droms, R., Woodyatt, J., and Y. Lee, "Dual- 787 Stack Lite Broadband Deployments Following IPv4 788 Exhaustion", RFC 6333, August 2011. 790 [RFC6598] Weil, J., Kuarsingh, V., Donley, C., Liljenstolpe, C., and 791 M. Azinger, "IANA-Reserved IPv4 Prefix for Shared Address 792 Space", BCP 153, RFC 6598, April 2012. 794 Authors' Addresses 796 Victor Kuarsingh (editor) 797 Rogers Communications 798 8200 Dixie Road 799 Brampton, Ontario L6T 0C1 800 Canada 802 Email: victor.kuarsingh@gmail.com 803 URI: http://www.rogers.com 805 John Cianfarani 806 Rogers Communications 807 8200 Dixie Road 808 Brampton, Ontario L6T 0C1 809 Canada 811 Email: john.cianfarani@rci.rogers.com 812 URI: http://www.rogers.com