idnits 2.17.1 draft-ietf-rtgwg-net2cloud-problem-statement-01.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 5 instances of too long lines in the document, the longest one being 48 characters in excess of 72. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (June 5, 2019) is 1788 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- == Unused Reference: 'ITU-T-X1036' is defined on line 654, but no explicit reference was found in the text Summary: 1 error (**), 0 flaws (~~), 2 warnings (==), 1 comment (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 1 Network Working Group L. Dunbar 2 Internet Draft A. Malis 3 Intended status: Informational Futurewei 4 Expires: Dec 2019 C. Jacquenet 5 Orange 6 M. Toy 7 Verizon 8 June 5, 2019 10 Seamless Interconnect Underlay to Cloud Overlay Problem Statement 11 draft-ietf-rtgwg-net2cloud-problem-statement-01 13 Abstract 15 This document describes the problems that enterprises face today 16 when connecting their branch offices to dynamic workloads in third 17 party data centers (a.k.a. Cloud DCs). 19 It examines some of the approaches interconnecting cloud DCs with 20 enterprises' on-premises DCs & branch offices. This document also 21 describes some of the (network) problems that many enterprises face 22 when they have workloads & applications & data split among hybrid 23 data centers, especially for those enterprises with multiple sites 24 that are already interconnected by VPNs (e.g., MPLS L2VPN/L3VPN). 26 Current operational problems are examined to determine whether there 27 is a need to improve existing protocols or whether a new protocol is 28 necessary to solve them. 30 Status of this Memo 32 This Internet-Draft is submitted in full conformance with the 33 provisions of BCP 78 and BCP 79. 35 Internet-Drafts are working documents of the Internet Engineering 36 Task Force (IETF), its areas, and its working groups. Note that 37 other groups may also distribute working documents as Internet- 38 Drafts. 40 Internet-Drafts are draft documents valid for a maximum of six 41 months and may be updated, replaced, or obsoleted by other documents 42 at any time. It is inappropriate to use Internet-Drafts as 43 reference material or to cite them other than as "work in progress." 45 The list of current Internet-Drafts can be accessed at 46 http://www.ietf.org/ietf/1id-abstracts.txt 48 The list of Internet-Draft Shadow Directories can be accessed at 49 http://www.ietf.org/shadow.html 51 This Internet-Draft will expire on December 5, 2019. 53 Copyright Notice 55 Copyright (c) 2019 IETF Trust and the persons identified as the 56 document authors. All rights reserved. 58 This document is subject to BCP 78 and the IETF Trust's Legal 59 Provisions Relating to IETF Documents 60 (http://trustee.ietf.org/license-info) in effect on the date of 61 publication of this document. Please review these documents 62 carefully, as they describe your rights and restrictions with 63 respect to this document. Code Components extracted from this 64 document must include Simplified BSD License text as described in 65 Section 4.e of the Trust Legal Provisions and are provided without 66 warranty as described in the Simplified BSD License. 68 Table of Contents 70 1. Introduction...................................................3 71 1.1. On the evolution of Cloud DC connectivity.................3 72 1.2. The role of SD-WAN techniques in Cloud DC connectivity....4 73 2. Definition of terms............................................4 74 3. Current Practices in Interconnecting Enterprise Sites with Cloud 75 DCs...............................................................5 76 3.1. Interconnect to Cloud DCs.................................5 77 3.2. Interconnect to Hybrid Cloud DCs..........................8 78 3.3. Connecting workloads among hybrid Cloud DCs...............8 79 4. Desired Properties for Networks that interconnect Hybrid Clouds9 80 5. Problems with MPLS-based VPNs extending to Hybrid Cloud DCs...10 81 6. Problem with using IPsec tunnels to Cloud DCs.................11 82 6.1. Complexity of multi-point any-to-any interconnection.....12 83 6.2. Poor performance over long distance......................12 84 6.3. Scaling Issues with IPsec Tunnels........................13 86 7. Problems of Using SD-WAN to connect to Cloud DCs..............13 87 7.1. SD-WAN among branch offices vs. interconnect to Cloud DCs13 88 8. End-to-End Security Concerns for Data Flows...................16 89 9. Requirements for Dynamic Cloud Data Center VPNs...............16 90 10. Security Considerations......................................17 91 Solution drafts resulting from this work will address security 92 concerns inherent to the solution(s), including both protocol 93 aspects and the importance (for example) of securing workloads in 94 cloud DCs and the use of secure interconnection mechanisms.......17 95 11. IANA Considerations..........................................17 96 12. References...................................................17 97 12.1. Normative References....................................17 98 12.2. Informative References..................................17 99 13. Acknowledgments..............................................18 101 1. Introduction 103 1.1. On the evolution of Cloud DC connectivity 105 The ever-increasing use of cloud applications for communication 106 services change the way corporate business works and shares 107 information. Such cloud applications use resources hosted in third 108 party DCs that also host services for other customers. 110 With the advent of widely available third party cloud DCs in diverse 111 geographic locations and the advancement of tools for monitoring and 112 predicting application behaviors, it is technically feasible for 113 enterprises to instantiate applications and workloads in locations 114 that are geographically closest to their end-users. Such proximity 115 improves end-to-end latency and overall user experience. Conversely, 116 an enterprise can easily shutdown applications and workloads 117 whenever end-users are in motion (thereby modifying the networking 118 connection of subsequently relocated applications and workloads). In 119 addition, an enterprise may wish to take advantage of more and more 120 business applications offered by third party private cloud DCs. 122 Most of those enterprise branch offices & on-premises data centers 123 are already connected via VPNs, such as MPLS-based L2VPNs and 124 L3VPNs. Then connecting to the cloud-hosted resources may not be 125 straightforward if the provider of the VPN service does not have 126 direct connections to the corresponding cloud DCs. Under those 127 circumstances, the enterprise can upgrade the CPEs deployed in its 128 various premises to utilize SD-WAN techniques to reach cloud 129 resources (without any assistance from the VPN service provider), or 130 wait for their VPN service provider to make new agreements with data 131 center providers to connect to the cloud resources. Either way has 132 additional infrastructure and operational costs. 134 In addition, it is an uptrend with more enterprises instantiating 135 their apps & workloads in different cloud DCs to maximize the 136 benefits of geographical proximity, elasticity and special features 137 offered by different cloud DCs. 139 1.2. The role of SD-WAN techniques in Cloud DC connectivity 141 This document discusses issues raised by the connection of 142 enterprise premises to third party data centers (a.k.a. Cloud DCs) 143 for reaching dynamic workloads. 145 SD-WAN, initially launched to maximize bandwidths between locations 146 by aggregating multiple paths managed by different service 147 providers, has expanded to include flexible, on-demand, application- 148 based connections established over any networks to access dynamic 149 workloads in Cloud DCs. 151 As a consequence, this document discusses the use of SD-WAN 152 techniques as a means to improve enterprise-to-cloud DC 153 connectivity. 155 2. Definition of terms 157 Cloud DC: Third party Data Centers that usually host applications 158 and workload owned by different organizations or 159 tenants. 161 Controller: Used interchangeably with SD-WAN controller to manage 162 SD-WAN overlay path creation/deletion and monitoring the 163 path conditions between two or more sites. 165 DSVPN: Dynamic Smart Virtual Private Network. DSVPN is a secure 166 network that exchanges data between sites without 167 needing to pass traffic through an organization's 168 headquarter virtual private network (VPN) server or 169 router. 171 Heterogeneous Cloud: applications & workloads split among Cloud DCs 172 owned & managed by different operators. 174 Hybrid Clouds: Hybrid Clouds (usually plural) refer to enterprises 175 using their own premises DCs in addition to Cloud 176 services provided by multiple cloud operators. For 177 example, an enterprise not only have applications 178 running in their own DCs, but also have applications 179 hosted in multiple third party cloud DCs ((AWS, Azure, 180 Google, Salesforces, SAP, etc). . ONUG also has a 181 notion of heterogeneous cloud, refers to enterprises 182 does not have its own DC, only uses services by 3rd 183 party cloud operators. 185 SD-WAN: Software Defined Wide Area Network. In this document, 186 "SD-WAN" refers to the solutions of pooling WAN 187 bandwidth from multiple underlay networks to get better 188 WAN bandwidth management, visibility & control. When the 189 underlay networks are private networks, traffic can 190 traverse without additional encryption; when the 191 underlay networks are public, such as Internet, some 192 traffic needs to be encrypted when traversing through 193 (depending on user provided policies). 195 VPC: Virtual Private Cloud. A service offered by Cloud DC 196 operators to allocate logically-isolated cloud 197 resources, including compute, networking and storage. 199 3. Current Practices in Interconnecting Enterprise Sites with Cloud DCs 201 3.1. Interconnect to Cloud DCs 203 Most Cloud operators offer some type of network gateway through 204 which an enterprise can reach their workloads hosted in the Cloud 205 DCs. For example, AWS (Amazon Web Services) offers the following 206 options to reach workloads in AWS Cloud DCs: 208 - Internet gateway for any external entities to reach the 209 workloads hosted in AWS Cloud DC via the Internet. 210 - Virtual gateway (vGW) where IPsec tunnels [RFC6071] are 211 established between an enterprise's own gateway and AWS vGW, so 212 that the communications between those gateways can be secured 213 from the underlay (which might be the public Internet). 214 - Direct Connect, which allows enterprises to purchase direct 215 connect from network service providers to get a private leased 216 line interconnecting the enterprises gateway(s) and the AWS 217 Direct Connect routers. Via Direct Connect, an AWS Transit 218 Gateway can be used to interconnect multiple VPCs in different 219 Availability Zones. 221 CPEs at one Enterprise branch office are connected to the Internet 222 to reach AWS's vGW via IPsec tunnels. Other ports of such CPEs are 223 connected to AWS DirectConnect via a private network (without any 224 encryption). 226 +------------------------+ 227 | ,---. ,---. | 228 | (TN-1 ) ( TN-2)| 229 | `-+-' +--+ `-+-' | 230 | +----|vR|-----+ | 231 | ++-+ | 232 | | +-+----+ 233 | | /Internet\ For External 234 | +-------+ Gateway +---------------------- 235 | \ / to reach via Internet 236 | +-+----+ 237 | | 238 +------------------------+ 240 +------------------------+ 241 | ,---. ,---. | 242 | (TN-1 ) ( TN-2)| 243 | `-+-' +--+ `-+-' | 244 | +----|vR|-----+ | 245 | ++-+ | 246 | | +-+----+ 247 | | / virtual\ For IPsec Tunnel 248 | +-------+ Gateway +---------------------- 249 | \ / termination 250 | +-+----+ 251 | | 252 +------------------------+ 254 +------------------------+ 255 | ,---. ,---. | 256 | (TN-1 ) ( TN-2)| 257 | `-+-' +--+ `-+-' | 258 | +----|vR|-----+ | 259 | ++-+ | 260 | | +-+----+ +------+ 261 | | / \ For Direct /customer\ 262 | +-------+ Gateway +----------+ gateway | 263 | \ / Connect \ / 264 | +-+----+ +------+ 265 | | 266 +------------------------+ 268 Figure 1: Examples of Cloud DC connections. 270 3.2. Interconnect to Hybrid Cloud DCs 272 According to Gartner, by 2020 "hybrid will be the most common usage 273 of the cloud" as more enterprises see the benefits of integrating 274 public and private cloud infrastructures. However, enabling the 275 growth of hybrid cloud deployments in the enterprise requires fast 276 and safe interconnection between public and private cloud services. 277 For an enterprise to connect to applications & workloads hosted in 278 multiple Cloud DCs, the enterprise can use IPsec tunnels established 279 over the Internet or a (virtualized) leased line service to connect 280 its on-premises gateways to each of the Cloud DC's gateways, virtual 281 routers instantiated in the Cloud DCs, or any other suitable design 282 (including a combination thereof). 284 Some enterprises prefer to instantiate their own virtual 285 CPEs/routers inside the Cloud DC to connect the workloads within the 286 Cloud DC. Then an overlay path is established between customer 287 gateways to the virtual CPEs/routers for reaching the workloads 288 inside the cloud DC. 290 3.3. Connecting workloads among hybrid Cloud DCs 292 There are multiple approaches to interconnect workloads among 293 different Cloud DCs: 295 - Utilize Cloud DC provided transit gateways, which usually does 296 not work if Cloud DCs are owned and managed by different Cloud 297 providers. 298 - Hairpin all the traffic through the customer gateway, which 299 creates additional transmission delay & incurs cost when 300 exiting Cloud DCs, or 301 - Establish direct tunnels among different VPCs (Virtual Private 302 Clouds) via client's own virtual routers instantiated within 303 Cloud DCs. DMVPN (Dynamic Multipoint Virtual Private Network) 304 or DSVPN (Dynamic Smart VPN) techniques can be used to 305 establish direct Multi-point-to-Point or multi-point-to multi- 306 point tunnels among those client's own virtual routers. 308 DMVPN & DSVPN use NHRP (Next Hop Resolution Protocol) [RFC2735] so 309 that spoke nodes can register their IP addresses & WAN ports with 310 the hub node. The IETF ION (Internetworking over NBMA (non-broadcast 311 multiple access) WG standardized NHRP for connection-oriented NBMA 312 network (such as ATM) network address resolution more than two 313 decades ago. 315 There are many differences between virtual routers in Public Cloud 316 DCs and the nodes in an NBMA network. NHRP & DSVPN are not cannot be 317 used for registering virtual routers in Cloud DCs unless an 318 extension of such protocols is developed for that purpose. Other 319 protocols such as BGP can be used, as described in [BGP-SDWAN]. 321 4. Desired Properties for Networks that interconnect Hybrid Clouds 322 The networks that interconnect hybrid cloud DCs must address the 323 following requirements: 324 - High availability at any time, whatever the duration of the 325 connection to the cloud DC. 326 Many enterprises include cloud infrastructures in their 327 disaster recovery strategy, e.g., by enforcing periodic backup 328 policies within the cloud, or by running backup applications in 329 the Cloud, etc. Therefore, the connection to the cloud DCs may 330 not be permanent, but rather needs to be on-demand. 332 - Global reachability from different geographical zones, thereby 333 facilitating the proximity of applications as a function of the 334 end users' location, to improve latency. 335 - Elasticity and mobility, to instantiate additional applications 336 at Cloud DCs when end-users' usages increase and shut down 337 applications at locations when there are fewer end-users. 338 Some enterprises have front-end web portals running in cloud 339 DCs and database servers in their on-premises DCs. Those Front- 340 end web portals need to be reachable from the public Internet. 341 The backend connection to the sensitive data in database 342 servers hosted in the on-premises DCs might need secure 343 connections. 345 - Scalable security management. IPsec is commonly used to 346 interconnect cloud gateways with CPEs deployed in the 347 enterprise premises. For enterprises with a large number or 348 branch offices, managing the IPsec's Security Associations 349 among many nodes can be very difficult. 351 5. Problems with MPLS-based VPNs extending to Hybrid Cloud DCs 353 Traditional MPLS-based VPNs have been widely deployed as an 354 effective way to support businesses and organizations that require 355 network performance and reliability. MPLS shifted the burden of 356 managing a VPN service from enterprises to service providers. The 357 CPEs attached to MPLS VPNs are also simpler and less expensive, 358 since they do not need to manage routes to remote sites; they simply 359 pass all outbound traffic to the MPLS VPN PEs to which the CPEs are 360 attached (albeit multi-homing scenarios require more processing 361 logic on CPEs). MPLS has addressed the problems of scale, 362 availability, and fast recovery from network faults, and 363 incorporated traffic-engineering capabilities. 365 However, traditional MPLS-based VPN solutions are sub-optimized for 366 connecting end-users to dynamic workloads/applications in cloud DCs 367 because: 369 - The Provider Edge (PE) nodes of the enterprise's VPNs might not 370 have direct connections to third party cloud DCs that are used 371 for hosting workloads with the goal of providing an easy access 372 to enterprises' end-users. 374 - It usually takes some time to deploy provider edge (PE) routers 375 at new locations. When enterprise's workloads are changed from 376 one cloud DC to another (i.e., removed from one DC and re- 377 instantiated to another location when demand changes), the 378 enterprise branch offices need to be connected to the new cloud 379 DC, but the network service provider might not have PEs located 380 at the new location. 382 One of the main drivers for moving workloads into the cloud is 383 the widely available cloud DCs at geographically diverse 384 locations, where apps can be instantiated so that they can be 385 as close to their end-users as possible. When the user base 386 changes, the applications may be migrated to a new cloud DC 387 location closest to the new user base. 389 - Most of the cloud DCs do not expose their internal networks, so 390 the MPLS-based VPNs can only reach Cloud DC's Gateways, not to 391 the workloads hosted inside. 393 - Many cloud DCs use an overlay to connect their gateways to the 394 workloads located inside the DC. There is currently no standard 395 that specifies the interworking between the Cloud Overlay and 396 the enterprise' existing underlay networks. One of the 397 characteristics of overlay networks is that some of the WAN 398 rd ports of the edge nodes connect to 3 party networks. There is 399 therefore a need to propagate WAN port information to remote 400 rd authorized peers in 3 party network domains in addition to 401 route propagation. Such an exchange cannot happen before 402 communication between peers is properly secured. 404 Another roadblock is the lack of a standard way to express and 405 enforce consistent security policies for workloads that not only use 406 virtual addresses, but in which are also very likely hosted in 407 different locations within the Cloud DC [RFC8192]. The current VPN 408 path computation and bandwidth allocation schemes may not be 409 flexible enough to address the need for enterprises to rapidly 410 connect to dynamically instantiated (or removed) workloads and 411 applications regardless of their location/nature (i.e., third party 412 cloud DCs). 414 6. Problem with using IPsec tunnels to Cloud DCs 415 As described in the previous section, many Cloud operators expose 416 their gateways for external entities (which can be enterprises 417 themselves) to directly establish IPsec tunnels. Enterprises can 418 also instantiate virtual routers within Cloud DCs to connect to 419 their on-premises devices via IPsec tunnels. If there is only one 420 enterprise location that needs to reach the Cloud DC, an IPsec 421 tunnel is a very convenient solution. 423 However, many medium-to-large enterprises usually have multiple 424 sites and multiple data centers. For workloads and apps hosted in 425 cloud DCs, multiple sites need to communicate securely with those 426 cloud workloads and apps. This section documents some of the issues 427 associated with using IPsec tunnels to connect enterprise premises 428 with cloud gateways. 430 6.1. Complexity of multi-point any-to-any interconnection 432 The dynamic workload instantiated in cloud DC needs to communicate 433 with multiple branch offices and on-premises data centers. Most 434 enterprises need multi-point interconnection among multiple 435 locations, which can be provided by means of MPLS L2/L3 VPNs. 437 Using IPsec overlay paths to connect all branches & on-premises data 438 centers to cloud DCs requires CPEs to manage routing among Cloud DCs 439 gateways and the CPEs located at other branch locations, which can 440 dramatically increase the complexity of the design, possibly at the 441 cost of jeopardizing the CPE performance. 443 The complexity of requiring CPEs to maintain routing among other 444 CPEs is one of the reasons why enterprises migrated from Frame Relay 445 based services to MPLS-based VPN services. 447 MPLS-based VPNs have their PEs directly connected to the CPEs. 448 Therefore, CPEs only need to forward all traffic to the directly 449 attached PEs, which are therefore responsible for enforcing the 450 routing policy within the corresponding VPNs. Even for multi-homed 451 CPEs, the CPEs only need to forward traffic among the directly 452 connected PEs. However, when using IPsec tunnels between CPEs and 453 Cloud DCs, the CPEs need to compute, select, establish and maintain 454 routes for traffic to be forwarded to Cloud DCs, to remote CPEs via 455 VPN, or directly. 457 6.2. Poor performance over long distance 459 When enterprise CPEs or gateways are far away from cloud DC gateways 460 or across country/continent boundaries, performance of IPsec tunnels 461 over the public Internet can be problematic and unpredictable. Even 462 though there are many monitoring tools available to measure delay 463 and various performance characteristics of the network, the 464 measurement for paths over the Internet is passive and past 465 measurements may not represent future performance. 467 Many cloud providers can replicate workloads in different available 468 zones. An App instantiated in a cloud DC closest to clients may have 469 to cooperate with another App (or its mirror image) in another 470 region or database server(s) in the on-premises DC. This kind of 471 coordination requires predicable networking behavior/performance 472 among those locations. 474 6.3. Scaling Issues with IPsec Tunnels 476 IPsec can achieve secure overlay connections between two locations 477 over any underlay network, e.g., between CPEs and Cloud DC Gateways. 479 If there is only one enterprise location connected to the cloud 480 gateway, a small number of IPsec tunnels can be configured on-demand 481 between the on-premises DC and the Cloud DC, which is an easy and 482 flexible solution. 484 However, for multiple enterprise locations to reach workloads hosted 485 in cloud DCs, the cloud DC gateway needs to maintain multiple IPsec 486 tunnels to all those locations (e.g., as a hub & spoke topology). 487 For a company with hundreds or thousands of locations, there could 488 be hundreds (or even thousands) of IPsec tunnels terminating at the 489 cloud DC gateway, which is not only very expensive (because Cloud 490 Operators usually charge their customers based on connections), but 491 can be very processing intensive for the gateway. Many cloud 492 operators only allow a limited number of (IPsec) tunnels & bandwidth 493 to each customer. Alternatively, you could use a solution like 494 group encryption where a single IPsec SA is necessary at the GW but 495 the drawback here is key distribution and maintenance of a key 496 server, etc. 498 7. Problems of Using SD-WAN to connect to Cloud DCs 499 SD-WAN can establish parallel paths over multiple underlay networks 500 between two locations on-demand, for example, to support the 501 connections established between two CPEs interconnected by a 502 traditional MPLS VPN ([RFC4364] or [RFC4664]) or by IPsec [RFC6071] 503 tunnels. 505 SD-WAN lets enterprises augment their current VPN network with cost- 506 effective, readily available Broadband Internet connectivity, 507 enabling some traffic offloading to paths over the Internet 508 according to differentiated, possibly application-based traffic 509 forwarding policies, or when the MPLS VPN connection between the two 510 locations is congested, or otherwise undesirable or unavailable. 512 7.1. SD-WAN among branch offices vs. interconnect to Cloud DCs 514 SD-WAN interconnection of branch offices is not as simple as it 515 appears. For an enterprise with multiple sites, using SD-WAN overlay 516 paths among sites requires each CPE to manage all the addresses that 517 local hosts have the potential to reach, i.e., map internal VPN 518 addresses to appropriate SD-WAN paths. This is similar to the 519 complexity of Frame Relay based VPNs, where each CPE needed to 520 maintain mesh routing for all destinations if they were to avoid an 521 extra hop through a hub router. Even though SD-WAN CPEs can get 522 assistance from a central controller (instead of running a routing 523 protocol) to resolve the mapping between destinations and SD-WAN 524 paths, SD-WAN CPEs are still responsible for routing table 525 maintenance as remote destinations change their attachments, e.g., 526 the dynamic workload in other DCs are de-commissioned or added. 528 Even though originally envisioned for interconnecting branch 529 offices, SD-WAN offers a very attractive way for enterprises to 530 connect to Cloud DCs. 532 The SD-WAN for interconnecting branch offices and the SD-WAN for 533 interconnecting to Cloud DCs have some differences: 535 - SD-WAN for interconnecting branch offices usually have two end- 536 points (e.g., CPEs) controlled by one entity (e.g., a 537 controller or management system operated by the enterprise). 538 - SD-WAN for Cloud DC interconnects may consider CPEs owned or 539 managed by the enterprise, while remote end-points are being 540 managed or controlled by Cloud DCs (For the ease of 541 description, let's call such CPEs asymmetrically-managed CPEs). 543 - Cloud DCs may have different entry points (or devices) with one 544 entry point that terminates a private direct connection (based 545 upon a leased line for example) and other entry points being 546 devices terminating the IPsec tunnels, as shown in Figure 2. 548 Therefore, the SD-WAN design becomes asymmetric. 549 +------------------------+ 550 | ,---. ,---. | 551 | (TN-1 ) ( TN-2)| 552 | `-+-' +---+ `-+-' | 553 | +----|vR1|----+ | 554 | ++--+ | 555 | | +-+----+ 556 | | /Internet\ One path via 557 | +-------+ Gateway +---------------------+ 558 | \ / Internet \ 559 | +-+----+ \ 560 +------------------------+ \ 561 \ 562 +------------------------+ native traffic \ 563 | ,---. ,---. | without encryption| 564 | (TN-3 ) ( TN-4)| | 565 | `-+-' +--+ `-+-' | | +------+ 566 | +----|vR|-----+ | +------+ CPE | 567 | ++-+ | | +------+ 568 | | +-+----+ | 569 | | / virtual\ One path via IPsec Tunnel | 570 | +-------+ Gateway +-------------------------- + 571 | \ / Encrypted traffic over| 572 | +-+----+ public network | 573 +------------------------+ | 574 | 575 +------------------------+ | 576 | ,---. ,---. | Native traffic | 577 | (TN-5 ) ( TN-6)| without encryption | 578 | `-+-' +--+ `-+-' | over secure network| 579 | +----|vR|-----+ | | 580 | ++-+ | | 581 | | +-+----+ +------+ | 582 | | / \ Via Direct /customer\ | 583 | +-------+ Gateway +----------+ gateway |-----+ 584 | \ / Connect \ / 585 | +-+----+ +------+ 586 +------------------------+ 588 Figure 2: Different Underlays to Reach Cloud DC 590 8. End-to-End Security Concerns for Data Flows 592 When IPsec tunnels established from enterprise on-premises CPEs 593 are terminated at the Cloud DC gateway where the workloads or 594 applications are hosted, some enterprises have concerns regarding 595 traffic to/from their workload being exposed to others behind the 596 data center gateway (e.g., exposed to other organizations that 597 have workloads in the same data center). 598 To ensure that traffic to/from workloads is not exposed to 599 unwanted entities, IPsec tunnels may go all the way to the 600 workload (servers, or VMs) within the DC. 602 9. Requirements for Dynamic Cloud Data Center VPNs 604 In order to address the aforementioned issues, any solution for 605 enterprise VPNs that includes connectivity to dynamic workloads or 606 applications in cloud data centers should satisfy a set of 607 requirements: 609 - The solution should allow enterprises to take advantage of the 610 current state-of-the-art in VPN technology, in both traditional 611 MPLS-based VPNs and IPsec-based VPNs (or any combination 612 thereof) that run over the public Internet. 613 - The solution should not require an enterprise to upgrade all 614 their existing CPEs. 615 - The solution should support scalable IPsec key management among 616 all nodes involved in DC interconnect schemes. 617 - The solution needs to support easy and fast, on-the-fly, VPN 618 connections to dynamic workloads and applications in third 619 party data centers, and easily allow these workloads to migrate 620 both within a data center and between data centers. 621 - Allow VPNs to provide bandwidth and other performance 622 guarantees. 623 - Be a cost-effective solution for enterprises to incorporate 624 dynamic cloud-based applications and workloads into their 625 existing VPN environment. 627 10. Security Considerations 629 The draft discusses security requirements as a part of the problem 630 space, particularly in sections 4, 5, and 8. 632 Solution drafts resulting from this work will address security 633 concerns inherent to the solution(s), including both protocol 634 aspects and the importance (for example) of securing workloads in 635 cloud DCs and the use of secure interconnection mechanisms. 637 11. IANA Considerations 639 This document requires no IANA actions. RFC Editor: Please remove 640 this section before publication. 642 12. References 644 12.1. Normative References 646 12.2. Informative References 648 [RFC2735] B. Fox, et al "NHRP Support for Virtual Private 649 networks". Dec. 1999. 651 [RFC8192] S. Hares, et al "Interface to Network Security Functions 652 (I2NSF) Problem Statement and Use Cases", July 2017 654 [ITU-T-X1036] ITU-T Recommendation X.1036, "Framework for creation, 655 storage, distribution and enforcement of policies for 656 network security", Nov 2007. 658 [RFC6071] S. Frankel and S. Krishnan, "IP Security (IPsec) and 659 Internet Key Exchange (IKE) Document Roadmap", Feb 2011. 661 [RFC4364] E. Rosen and Y. Rekhter, "BGP/MPLS IP Virtual Private 662 Networks (VPNs)", Feb 2006 664 [RFC4664] L. Andersson and E. Rosen, "Framework for Layer 2 Virtual 665 Private Networks (L2VPNs)", Sept 2006. 667 [BGP-SDWAN] L. Dunbar, et al. "BGP Extension for SDWAN Overlay 668 Networks", draft-dunbar-idr-bgp-sdwan-overlay-ext-03, 669 work-in-progress, Nov 2018. 671 13. Acknowledgments 673 Many thanks to Ignas Bagdonas, Michael Huang, Liu Yuan Jiao, 674 Katherine Zhao, and Jim Guichard for the discussion and 675 contributions. 677 Authors' Addresses 679 Linda Dunbar 680 Futurewei 681 Email: Linda.Dunbar@huawei.com 683 Andrew G. Malis 684 Futurewei 685 Email: agmalis@gmail.com 687 Christian Jacquenet 688 Orange 689 Rennes, 35000 690 France 691 Email: Christian.jacquenet@orange.com 693 Mehmet Toy 694 Verizon 695 One Verizon Way 696 Basking Ridge, NJ 07920 697 Email: mehmet.toy@verizon.com