idnits 2.17.1 draft-dm-net2cloud-problem-statement-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 3 instances of too long lines in the document, the longest one being 8 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 (February 6, 2019) is 1896 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- == Unused Reference: 'ITU-T-X1036' is defined on line 631, 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 Huawei 4 Expires: July 2019 C. Jacquenet 5 Orange 6 M. Toy 7 Verizon 8 February 6, 2019 10 Seamless Interconnect Underlay to Cloud Overlay Problem Statement 11 draft-dm-net2cloud-problem-statement-07 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 August 6, 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 2. Definition of terms............................................4 72 3. Current Practices in Interconnecting Enterprise Sites with Cloud 73 DCs...............................................................5 74 3.1. Interconnect to Cloud DCs.................................5 75 3.2. Interconnect to Hybrid Cloud DCs..........................7 76 3.3. Connecting workloads among hybrid Cloud DCs...............7 77 4. Desired Properties for Networks that interconnect Hybrid Clouds8 78 5. Problems with MPLS-based VPNs extending to Hybrid Cloud DCs....9 79 6. Problem with using IPsec tunnels to Cloud DCs.................10 80 6.1. Complexity of multi-point any-to-any interconnection.....10 81 6.2. Poor performance over long distance......................11 82 6.3. Scaling Issues with IPsec Tunnels........................11 83 7. Problems of Using SD-WAN to connect to Cloud DCs..............12 84 7.1. SD-WAN among branch offices vs. interconnect to Cloud DCs12 86 8. End-to-End Security Concerns for Data Flows...................15 87 9. Requirements for Dynamic Cloud Data Center VPNs...............15 88 10. Security Considerations......................................16 89 Solution drafts resulting from this work will address security 90 concerns inherent to the solution(s), including both protocol 91 aspects and the importance (for example) of securing workloads in 92 cloud DCs and the use of secure interconnection mechanisms.......16 93 IANA Considerations..............................................16 94 11. References...................................................16 95 11.1. Normative References....................................16 96 11.2. Informative References..................................16 97 12. Acknowledgments..............................................17 99 1. Introduction 101 The ever-increasing use of cloud applications for communication 102 services change the way corporate business works and shares 103 information. Such cloud applications use resources hosted in third 104 party DCs that also host services for other customers. 106 With the advent of widely available third party cloud DCs in diverse 107 geographic locations and the advancement of tools for monitoring and 108 predicting application behaviors, it is technically feasible for 109 enterprises to instantiate applications and workloads in locations 110 that are geographically closest to their end-users. Such proximity 111 improves end-to-end latency and overall user experience. Conversely, 112 an enterprise can easily shutdown applications and workloads 113 whenever end-users are in motion (thereby modifying the networking 114 connection of subsequently relocated applications and workloads). In 115 addition, an enterprise may wish to take advantage of more and more 116 business applications offered by third party private cloud DCs. 118 Most of those enterprise branch offices & on-premises data centers 119 are already connected via VPNs, such as MPLS-based L2VPNs and 120 L3VPNs. Then connecting to the cloud-hosted resources may not be 121 straightforward if the provider of the VPN service does not have 122 direct connections to the corresponding cloud DCs. Under those 123 circumstances, the enterprise can upgrade the CPEs deployed in its 124 various premises to utilize SD-WAN techniques to reach cloud 125 resources (without any assistance from the VPN service provider), or 126 wait for their VPN service provider to make new agreements with data 127 center providers to connect to the cloud resources. Either way has 128 additional infrastructure and operational costs. 130 In addition, it is an uptrend with more enterprises instantiating 131 their apps & workloads in different cloud DCs to maximize the 132 benefits of geographical proximity, elasticity and special features 133 offered by different cloud DCs. 135 2. Definition of terms 137 Cloud DC: Third party Data Centers that usually host applications 138 and workload owned by different organizations or 139 tenants. 141 Controller: Used interchangeably with SD-WAN controller to manage 142 SD-WAN overlay path creation/deletion and monitoring the 143 path conditions between two or more sites. 145 DSVPN: Dynamic Smart Virtual Private Network. DSVPN is a secure 146 network that exchanges data between sites without 147 needing to pass traffic through an organization's 148 headquarter virtual private network (VPN) server or 149 router. 151 Heterogeneous Cloud: applications & workloads split among Cloud DCs 152 owned & managed by different operators. 154 Hybrid Clouds: Hybrid Clouds (usually plural) refer to enterprises 155 using their own premises DCs in addition to Cloud 156 services provided by multiple cloud operators. For 157 example, an enterprise not only have applications 158 running in their own DCs, but also have applications 159 hosted in multiple third party cloud DCs ((AWS, Azure, 160 Google, Salesforces, SAP, etc). . ONUG also has a 161 notion of heterogeneous cloud, refers to enterprises 162 does not have its own DC, only uses services by 3rd 163 party cloud operators. 165 SD-WAN: Software Defined Wide Area Network. In this document, 166 "SD-WAN" refers to the solutions specified by ONUG (Open 167 Network User Group), https://www.onug.net/software- 168 defined-wide-area-network-sd-wan/, which is about 169 pooling WAN bandwidth from multiple underlay networks to 170 get better WAN bandwidth management, visibility & 171 control. When the underlay networks are private 172 networks, traffic can traverse without additional 173 encryption; when the underlay networks are public, such 174 as Internet, some traffic needs to be encrypted when 175 traversing through (depending on user provided 176 policies). 178 VPC: Virtual Private Cloud. A service offered by Cloud DC 179 operators to allocate logically-isolated cloud 180 resources, including compute, networking and storage. 182 3. Current Practices in Interconnecting Enterprise Sites with Cloud DCs 184 3.1. Interconnect to Cloud DCs 186 Most Cloud operators offer some type of network gateway through 187 which an enterprise can reach their workloads hosted in the Cloud 188 DCs. For example, AWS (Amazon Web Services) offers the following 189 options to reach workloads in AWS Cloud DCs: 191 - Internet gateway for any external entities to reach the 192 workloads hosted in AWS Cloud DC via the Internet. 193 - Virtual gateway (vGW) where IPsec tunnels [RFC6071] are 194 established between an enterprise's own gateway and AWS vGW, so 195 that the communications between those gateways can be secured 196 from the underlay (which might be the public Internet). 197 - Direct Connect, which allows enterprises to purchase direct 198 connect from network service providers to get a private leased 199 line interconnecting the enterprises gateway(s) and the AWS 200 Direct Connect routers. Via Direct Connect, an AWS Transit 201 Gateway can be used to interconnect multiple VPCs in different 202 Availability Zones. 204 CPEs at one Enterprise branch office are connected to the Internet 205 to reach AWS's vGW via IPsec tunnels. Other ports of such CPEs are 206 connected to AWS DirectConnect via a private network (without any 207 encryption). 209 +------------------------+ 210 | ,---. ,---. | 211 | (TN-1 ) ( TN-2)| 212 | `-+-' +--+ `-+-' | 213 | +----|vR|-----+ | 214 | ++-+ | 215 | | +-+----+ 216 | | /Internet\ For External 217 | +-------+ Gateway +---------------------- 218 | \ / to reach via Internet 219 | +-+----+ 220 | | 221 +------------------------+ 223 +------------------------+ 224 | ,---. ,---. | 225 | (TN-1 ) ( TN-2)| 226 | `-+-' +--+ `-+-' | 227 | +----|vR|-----+ | 228 | ++-+ | 229 | | +-+----+ 230 | | / virtual\ For IPsec Tunnel 231 | +-------+ Gateway +---------------------- 232 | \ / termination 233 | +-+----+ 234 | | 235 +------------------------+ 237 +------------------------+ 238 | ,---. ,---. | 239 | (TN-1 ) ( TN-2)| 240 | `-+-' +--+ `-+-' | 241 | +----|vR|-----+ | 242 | ++-+ | 243 | | +-+----+ +------+ 244 | | / \ For Direct /customer\ 245 | +-------+ Gateway +----------+ gateway | 246 | \ / Connect \ / 247 | +-+----+ +------+ 248 | | 249 +------------------------+ 251 Figure 1: Examples of Cloud DC connections. 253 3.2. Interconnect to Hybrid Cloud DCs 255 According to Gartner, by 2020 "hybrid will be the most common usage 256 of the cloud" as more enterprises see the benefits of integrating 257 public and private cloud infrastructures. However, enabling the 258 growth of hybrid cloud deployments in the enterprise requires fast 259 and safe interconnection between public and private cloud services. 260 For an enterprise to connect to applications & workloads hosted in 261 multiple Cloud DCs, the enterprise can use IPsec tunnels established 262 over the Internet or a (virtualized) leased line service to connect 263 its on-premises gateways to each of the Cloud DC's gateways, virtual 264 routers instantiated in the Cloud DCs, or any other suitable design 265 (including a combination thereof). 267 Some enterprises prefer to instantiate their own virtual 268 CPEs/routers inside the Cloud DC to connect the workloads within the 269 Cloud DC. Then an overlay path is established between customer 270 gateways to the virtual CPEs/routers for reaching the workloads 271 inside the cloud DC. 273 3.3. Connecting workloads among hybrid Cloud DCs 275 There are multiple approaches to interconnect workloads among 276 different Cloud DCs: 278 - Utilize Cloud DC provided transit gateways, which usually does 279 not work if Cloud DCs are owned and managed by different Cloud 280 providers. 281 - Hairpin all the traffic through the customer gateway, which 282 creates additional transmission delay & incurs cost when 283 exiting Cloud DCs, or 284 - Establish direct tunnels among different VPCs (Virtual Private 285 Clouds) via client's own virtual routers instantiated within 286 Cloud DCs. DMVPN (Dynamic Multipoint Virtual Private Network) 287 or DSVPN (Dynamic Smart VPN) techniques can be used to 288 establish direct Multi-point-to-Point or multi-point-to multi- 289 point tunnels among those client's own virtual routers. 291 DMVPN & DSVPN use NHRP (Next Hop Resolution Protocol) [RFC2735] so 292 that spoke nodes can register their IP addresses & WAN ports with 293 the hub node. The IETF ION (Internetworking over NBMA (non-broadcast 294 multiple access) WG standardized NHRP for connection-oriented NBMA 295 network (such as ATM) network address resolution more than two 296 decades ago. 298 There are many differences between virtual routers in Public Cloud 299 DCs and the nodes in an NBMA network. NHRP & DSVPN are not cannot be 300 used for registering virtual routers in Cloud DCs unless an 301 extension of such protocols is developed for that purpose. Other 302 protocols such as BGP can be used, as described in [BGP-SDWAN]. 304 4. Desired Properties for Networks that interconnect Hybrid Clouds 305 The networks that interconnect hybrid cloud DCs must address the 306 following requirements: 307 - High availability at any time, whatever the duration of the 308 connection to the cloud DC. 309 Many enterprises include cloud infrastructures in their 310 disaster recovery strategy, e.g., by enforcing periodic backup 311 policies within the cloud, or by running backup applications in 312 the Cloud, etc. Therefore, the connection to the cloud DCs may 313 not be permanent, but rather needs to be on-demand. 315 - Global reachability from different geographical zones, thereby 316 facilitating the proximity of applications as a function of the 317 end users' location, to improve latency. 318 - Elasticity and mobility, to instantiate additional applications 319 at Cloud DCs when end-users' usages increase and shut down 320 applications at locations when there are fewer end-users. 321 Some enterprises have front-end web portals running in cloud 322 DCs and database servers in their on-premises DCs. Those Front- 323 end web portals need to be reachable from the public Internet. 324 The backend connection to the sensitive data in database 325 servers hosted in the on-premises DCs might need secure 326 connections. 328 - Scalable security management. IPsec is commonly used to 329 interconnect cloud gateways with CPEs deployed in the 330 enterprise premises. For enterprises with a large number or 331 branch offices, managing the IPsec's Security Associations 332 among many nodes can be very difficult. 334 5. Problems with MPLS-based VPNs extending to Hybrid Cloud DCs 336 Traditional MPLS-based VPNs have been widely deployed as an 337 effective way to support businesses and organizations that require 338 network performance and reliability. MPLS shifted the burden of 339 managing a VPN service from enterprises to service providers. The 340 CPEs attached to MPLS VPNs are also simpler and less expensive, 341 since they do not need to manage routes to remote sites; they simply 342 pass all outbound traffic to the MPLS VPN PEs to which the CPEs are 343 attached (albeit multi-homing scenarios require more processing 344 logic on CPEs). MPLS has addressed the problems of scale, 345 availability, and fast recovery from network faults, and 346 incorporated traffic-engineering capabilities. 348 However, traditional MPLS-based VPN solutions are sub-optimized for 349 connecting end-users to dynamic workloads/applications in cloud DCs 350 because: 352 - The Provider Edge (PE) nodes of the enterprise's VPNs might not 353 have direct connections to third party cloud DCs that are used 354 for hosting workloads with the goal of providing an easy access 355 to enterprises' end-users. 357 - It usually takes some time to deploy provider edge (PE) routers 358 at new locations. When enterprise's workloads are changed from 359 one cloud DC to another (i.e., removed from one DC and re- 360 instantiated to another location when demand changes), the 361 enterprise branch offices need to be connected to the new cloud 362 DC, but the network service provider might not have PEs located 363 at the new location. 365 One of the main drivers for moving workloads into the cloud is 366 the widely available cloud DCs at geographically diverse 367 locations, where apps can be instantiated so that they can be 368 as close to their end-users as possible. When the user base 369 changes, the applications may be migrated to a new cloud DC 370 location closest to the new user base. 372 - Most of the cloud DCs do not expose their internal networks, so 373 the MPLS-based VPNs can only reach Cloud DC's Gateways, not to 374 the workloads hosted inside. 376 - Many cloud DCs use an overlay to connect their gateways to the 377 workloads located inside the DC. There has not been any 378 standard to address the interworking between the Cloud Overlay 379 and the enterprise' existing underlay networks. 381 Another roadblock is the lack of a standard way to express and 382 enforce consistent security policies for workloads that not only use 383 virtual addresses, but in which are also very likely hosted in 384 different locations within the Cloud DC [RFC8192]. The current VPN 385 path computation and bandwidth allocation schemes may not be 386 flexible enough to address the need for enterprises to rapidly 387 connect to dynamically instantiated (or removed) workloads and 388 applications regardless of their location/nature (i.e., third party 389 cloud DCs). 391 6. Problem with using IPsec tunnels to Cloud DCs 392 As described in the previous section, many Cloud operators expose 393 their gateways for external entities (which can be enterprises 394 themselves) to directly establish IPsec tunnels. Enterprises can 395 also instantiate virtual routers within Cloud DCs to connect to 396 their on-premises devices via IPsec tunnels. If there is only one 397 enterprise location that needs to reach the Cloud DC, an IPsec 398 tunnel is a very convenient solution. 400 However, many medium-to-large enterprises usually have multiple 401 sites and multiple data centers. For workloads and apps hosted in 402 cloud DCs, multiple sites need to communicate securely with those 403 cloud workloads and apps. This section documents some of the issues 404 associated with using IPsec tunnels to connect enterprise premises 405 with cloud gateways. 407 6.1. Complexity of multi-point any-to-any interconnection 409 The dynamic workload instantiated in cloud DC needs to communicate 410 with multiple branch offices and on-premises data centers. Most 411 enterprises need multi-point interconnection among multiple 412 locations, which can be provided by means of MPLS L2/L3 VPNs. 414 Using IPsec overlay paths to connect all branches & on-premises data 415 centers to cloud DCs requires CPEs to manage routing among Cloud DCs 416 gateways and the CPEs located at other branch locations, which can 417 dramatically increase the complexity of the design, possibly at the 418 cost of jeopardizing the CPE performance. 420 The complexity of requiring CPEs to maintain routing among other 421 CPEs is one of the reasons why enterprises migrated from Frame Relay 422 based services to MPLS-based VPN services. 424 MPLS-based VPNs have their PEs directly connected to the CPEs. 425 Therefore, CPEs only need to forward all traffic to the directly 426 attached PEs, which are therefore responsible for enforcing the 427 routing policy within the corresponding VPNs. Even for multi-homed 428 CPEs, the CPEs only need to forward traffic among the directly 429 connected PEs. However, when using IPsec tunnels between CPEs and 430 Cloud DCs, the CPEs need to compute, select, establish and maintain 431 routes for traffic to be forwarded to Cloud DCs, to remote CPEs via 432 VPN, or directly. 434 6.2. Poor performance over long distance 436 When enterprise CPEs or gateways are far away from cloud DC gateways 437 or across country/continent boundaries, performance of IPsec tunnels 438 over the public Internet can be problematic and unpredictable. Even 439 though there are many monitoring tools available to measure delay 440 and various performance characteristics of the network, the 441 measurement for paths over the Internet is passive and past 442 measurements may not represent future performance. 444 Many cloud providers can replicate workloads in different available 445 zones. An App instantiated in a cloud DC closest to clients may have 446 to cooperate with another App (or its mirror image) in another 447 region or database server(s) in the on-premises DC. This kind of 448 coordination requires predicable networking behavior/performance 449 among those locations. 451 6.3. Scaling Issues with IPsec Tunnels 453 IPsec can achieve secure overlay connections between two locations 454 over any underlay network, e.g., between CPEs and Cloud DC Gateways. 456 If there is only one enterprise location connected to the cloud 457 gateway, a small number of IPsec tunnels can be configured on-demand 458 between the on-premises DC and the Cloud DC, which is an easy and 459 flexible solution. 461 However, for multiple enterprise locations to reach workloads hosted 462 in cloud DCs, the cloud DC gateway needs to maintain multiple IPsec 463 tunnels to all those locations (e.g., as a hub & spoke topology). 464 For a company with hundreds or thousands of locations, there could 465 be hundreds (or even thousands) of IPsec tunnels terminating at the 466 cloud DC gateway, which is not only very expensive (because Cloud 467 Operators usually charge their customers based on connections), but 468 can be very processing intensive for the gateway. Many cloud 469 operators only allow a limited number of (IPsec) tunnels & bandwidth 470 to each customer. Alternatively, you could use a solution like 471 group encryption where a single IPsec SA is necessary at the GW but 472 the drawback here is key distribution and maintenance of a key 473 server, etc. 475 7. Problems of Using SD-WAN to connect to Cloud DCs 476 SD-WAN can establish parallel paths over multiple underlay networks 477 between two locations on-demand, for example, to support the 478 connections established between two CPEs interconnected by a 479 traditional MPLS VPN ([RFC4364] or [RFC4664]) or by IPsec [RFC6071] 480 tunnels. 482 SD-WAN lets enterprises augment their current VPN network with cost- 483 effective, readily available Broadband Internet connectivity, 484 enabling some traffic offloading to paths over the Internet 485 according to differentiated, possibly application-based traffic 486 forwarding policies, or when the MPLS VPN connection between the two 487 locations is congested, or otherwise undesirable or unavailable. 489 7.1. SD-WAN among branch offices vs. interconnect to Cloud DCs 491 SD-WAN interconnection of branch offices is not as simple as it 492 appears. For an enterprise with multiple sites, using SD-WAN overlay 493 paths among sites requires each CPE to manage all the addresses that 494 local hosts have the potential to reach, i.e., map internal VPN 495 addresses to appropriate SD-WAN paths. This is similar to the 496 complexity of Frame Relay based VPNs, where each CPE needed to 497 maintain mesh routing for all destinations if they were to avoid an 498 extra hop through a hub router. Even though SD-WAN CPEs can get 499 assistance from a central controller (instead of running a routing 500 protocol) to resolve the mapping between destinations and SD-WAN 501 paths, SD-WAN CPEs are still responsible for routing table 502 maintenance as remote destinations change their attachments, e.g., 503 the dynamic workload in other DCs are de-commissioned or added. 505 Even though originally envisioned for interconnecting branch 506 offices, SD-WAN offers a very attractive way for enterprises to 507 connect to Cloud DCs. 509 The SD-WAN for interconnecting branch offices and the SD-WAN for 510 interconnecting to Cloud DCs have some differences: 512 - SD-WAN for interconnecting branch offices usually have two end- 513 points (e.g., CPEs) controlled by one entity (e.g., a 514 controller or management system operated by the enterprise). 515 - SD-WAN for Cloud DC interconnects may consider CPEs owned or 516 managed by the enterprise, while remote end-points are being 517 managed or controlled by Cloud DCs (For the ease of 518 description, let's call such CPEs asymmetrically-managed CPEs). 520 - Cloud DCs may have different entry points (or devices) with one 521 entry point that terminates a private direct connection (based 522 upon a leased line for example) and other entry points being 523 devices terminating the IPsec tunnels, as shown in Figure 2. 525 Therefore, the SD-WAN design becomes asymmetric. 526 +------------------------+ 527 | ,---. ,---. | 528 | (TN-1 ) ( TN-2)| 529 | `-+-' +---+ `-+-' | 530 | +----|vR1|----+ | 531 | ++--+ | 532 | | +-+----+ 533 | | /Internet\ One path via 534 | +-------+ Gateway +---------------------+ 535 | \ / Internet \ 536 | +-+----+ \ 537 +------------------------+ \ 538 \ 539 +------------------------+ native traffic \ 540 | ,---. ,---. | without encryption| 541 | (TN-3 ) ( TN-4)| | 542 | `-+-' +--+ `-+-' | | +------+ 543 | +----|vR|-----+ | +------+ CPE | 544 | ++-+ | | +------+ 545 | | +-+----+ | 546 | | / virtual\ One path via IPsec Tunnel | 547 | +-------+ Gateway +-------------------------- + 548 | \ / Encrypted traffic over| 549 | +-+----+ public network | 550 +------------------------+ | 551 | 552 +------------------------+ | 553 | ,---. ,---. | Native traffic | 554 | (TN-5 ) ( TN-6)| without encryption | 555 | `-+-' +--+ `-+-' | over secure network| 556 | +----|vR|-----+ | | 557 | ++-+ | | 558 | | +-+----+ +------+ | 559 | | / \ Via Direct /customer\ | 560 | +-------+ Gateway +----------+ gateway |-----+ 561 | \ / Connect \ / 562 | +-+----+ +------+ 563 +------------------------+ 565 Figure 2: Different Underlays to Reach Cloud DC 567 8. End-to-End Security Concerns for Data Flows 569 When IPsec tunnels established from enterprise on-premises CPEs 570 are terminated at the Cloud DC gateway where the workloads or 571 applications are hosted, some enterprises have concerns regarding 572 traffic to/from their workload being exposed to others behind the 573 data center gateway (e.g., exposed to other organizations that 574 have workloads in the same data center). 575 To ensure that traffic to/from workloads is not exposed to 576 unwanted entities, IPsec tunnels may go all the way to the 577 workload (servers, or VMs) within the DC. 579 9. Requirements for Dynamic Cloud Data Center VPNs 581 In order to address the aforementioned issues, any solution for 582 enterprise VPNs that includes connectivity to dynamic workloads or 583 applications in cloud data centers should satisfy a set of 584 requirements: 586 - The solution should allow enterprises to take advantage of the 587 current state-of-the-art in VPN technology, in both traditional 588 MPLS-based VPNs and IPsec-based VPNs (or any combination 589 thereof) that run over the public Internet. 590 - The solution should not require an enterprise to upgrade all 591 their existing CPEs. 592 - The solution should support scalable IPsec key management among 593 all nodes involved in DC interconnect schemes. 594 - The solution needs to support easy and fast, on-the-fly, VPN 595 connections to dynamic workloads and applications in third 596 party data centers, and easily allow these workloads to migrate 597 both within a data center and between data centers. 598 - Allow VPNs to provide bandwidth and other performance 599 guarantees. 600 - Be a cost-effective solution for enterprises to incorporate 601 dynamic cloud-based applications and workloads into their 602 existing VPN environment. 604 10. Security Considerations 606 The draft discusses security requirements as a part of the problem 607 space, particularly in sections 4, 5, and 8. 609 Solution drafts resulting from this work will address security 610 concerns inherent to the solution(s), including both protocol 611 aspects and the importance (for example) of securing workloads in 612 cloud DCs and the use of secure interconnection mechanisms. 614 IANA Considerations 616 This document requires no IANA actions. RFC Editor: Please remove 617 this section before publication. 619 11. References 621 11.1. Normative References 623 11.2. Informative References 625 [RFC2735] B. Fox, et al "NHRP Support for Virtual Private 626 networks". Dec. 1999. 628 [RFC8192] S. Hares, et al "Interface to Network Security Functions 629 (I2NSF) Problem Statement and Use Cases", July 2017 631 [ITU-T-X1036] ITU-T Recommendation X.1036, "Framework for creation, 632 storage, distribution and enforcement of policies for 633 network security", Nov 2007. 635 [RFC6071] S. Frankel and S. Krishnan, "IP Security (IPsec) and 636 Internet Key Exchange (IKE) Document Roadmap", Feb 2011. 638 [RFC4364] E. Rosen and Y. Rekhter, "BGP/MPLS IP Virtual Private 639 Networks (VPNs)", Feb 2006 641 [RFC4664] L. Andersson and E. Rosen, "Framework for Layer 2 Virtual 642 Private Networks (L2VPNs)", Sept 2006. 644 [BGP-SDWAN] L. Dunbar, et al. "BGP Extension for SDWAN Overlay 645 Networks", draft-dunbar-idr-bgp-sdwan-overlay-ext-03, 646 work-in-progress, Nov 2018. 648 12. Acknowledgments 650 Many thanks to Ignas Bagdonas, Michael Huang, Liu Yuan Jiao, 651 Katherine Zhao, and Jim Guichard for the discussion and 652 contributions. 654 Authors' Addresses 656 Linda Dunbar 657 Huawei 658 Email: Linda.Dunbar@huawei.com 660 Andrew G. Malis 661 Huawei 662 Email: agmalis@gmail.com 664 Christian Jacquenet 665 Orange 666 Rennes, 35000 667 France 668 Email: Christian.jacquenet@orange.com 670 Mehmet Toy 671 Verizon 672 One Verizon Way 673 Basking Ridge, NJ 07920 674 Email: mehmet.toy@verizon.com