idnits 2.17.1 draft-dm-net2cloud-problem-statement-06.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 (January 25, 2019) is 1919 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- == Unused Reference: 'ITU-T-X1036' is defined on line 634, 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 January 25, 2019 10 Seamless Interconnect Underlay to Cloud Overlay Problem Statement 11 draft-dm-net2cloud-problem-statement-06 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 This Internet-Draft is submitted in full conformance with the 36 provisions of BCP 78 and BCP 79. This document may not be modified, 37 and derivative works of it may not be created, except to publish it 38 as an RFC and to translate it into languages other than English. 40 Internet-Drafts are working documents of the Internet Engineering 41 Task Force (IETF), its areas, and its working groups. Note that 42 other groups may also distribute working documents as Internet- 43 Drafts. 45 Internet-Drafts are draft documents valid for a maximum of six 46 months and may be updated, replaced, or obsoleted by other documents 47 at any time. It is inappropriate to use Internet-Drafts as 48 reference material or to cite them other than as "work in progress." 50 The list of current Internet-Drafts can be accessed at 51 http://www.ietf.org/ietf/1id-abstracts.txt 53 The list of Internet-Draft Shadow Directories can be accessed at 54 http://www.ietf.org/shadow.html 56 This Internet-Draft will expire on July 25, 2019. 58 Copyright Notice 60 Copyright (c) 2019 IETF Trust and the persons identified as the 61 document authors. All rights reserved. 63 This document is subject to BCP 78 and the IETF Trust's Legal 64 Provisions Relating to IETF Documents 65 (http://trustee.ietf.org/license-info) in effect on the date of 66 publication of this document. Please review these documents 67 carefully, as they describe your rights and restrictions with 68 respect to this document. Code Components extracted from this 69 document must include Simplified BSD License text as described in 70 Section 4.e of the Trust Legal Provisions and are provided without 71 warranty as described in the Simplified BSD License. 73 Table of Contents 75 1. Introduction...................................................3 76 2. Definition of terms............................................4 77 3. Current Practices in Interconnecting Enterprise Sites with Cloud 78 DCs...............................................................5 79 3.1. Interconnect to Cloud DCs.................................5 80 3.2. Interconnect to Hybrid Cloud DCs..........................7 81 3.3. Connecting workloads among hybrid Cloud DCs...............7 82 4. Desired Properties for Networks that interconnect Hybrid Clouds8 83 5. Problems with MPLS-based VPNs extending to Hybrid Cloud DCs....9 84 6. Problem with using IPsec tunnels to Cloud DCs.................10 85 6.1. Complexity of multi-point any-to-any interconnection.....10 86 6.2. Poor performance over long distance......................11 87 6.3. Scaling Issues with IPsec Tunnels........................11 88 7. Problems of Using SD-WAN to connect to Cloud DCs..............12 89 7.1. SD-WAN among branch offices vs. interconnect to Cloud DCs12 90 8. End-to-End Security Concerns for Data Flows...................15 91 9. Requirements for Dynamic Cloud Data Center VPNs...............15 92 10. Security Considerations......................................16 93 Solution drafts resulting from this work will address security 94 concerns inherent to the solution(s), including both protocol 95 aspects and the importance (for example) of securing workloads in 96 cloud DCs and the use of secure interconnection mechanisms.......16 97 IANA Considerations..............................................16 98 11. References...................................................16 99 11.1. Normative References....................................16 100 11.2. Informative References..................................16 101 12. Acknowledgments..............................................17 103 1. Introduction 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 2. Definition of terms 141 Cloud DC: Third party Data Centers that usually host applications 142 and workload owned by different organizations or 143 tenants. 145 Controller: Used interchangeably with SD-WAN controller to manage 146 SD-WAN overlay path creation/deletion and monitoring the 147 path conditions between two or more sites. 149 DSVPN: Dynamic Smart Virtual Private Network. DSVPN is a secure 150 network that exchanges data between sites without 151 needing to pass traffic through an organization's 152 headquarter virtual private network (VPN) server or 153 router. 155 Heterogeneous Cloud: applications & workloads split among Cloud DCs 156 owned & managed by different operators. 158 Hybrid Clouds: Hybrid Clouds (usually plural) refer to enterprises 159 using their own premises DCs in addition to Cloud 160 services provided by multiple cloud operators. For 161 example, an enterprise not only have applications 162 running in their own DCs, but also have applications 163 hosted in multiple third party cloud DCs ((AWS, Azure, 164 Google, Salesforces, SAP, etc). . ONUG also has a 165 notion of heterogeneous cloud, refers to enterprises 166 does not have its own DC, only uses services by 3rd 167 party cloud operators. 169 SD-WAN: Software Defined Wide Area Network. In this document, 170 "SD-WAN" refers to the solutions specified by ONUG (Open 171 Network User Group), https://www.onug.net/software- 172 defined-wide-area-network-sd-wan/, which is about 173 pooling WAN bandwidth from multiple underlay networks to 174 get better WAN bandwidth management, visibility & 175 control. When the underlay networks are private 176 networks, traffic can traverse without additional 177 encryption; when the underlay networks are public, such 178 as Internet, some traffic needs to be encrypted when 179 traversing through (depending on user provided 180 policies). 182 VPC: Virtual Private Cloud. A service offered by Cloud DC 183 operators to allocate logically-isolated cloud 184 resources, including compute, networking and storage. 186 3. Current Practices in Interconnecting Enterprise Sites with Cloud DCs 188 3.1. Interconnect to Cloud DCs 190 Most Cloud operators offer some type of network gateway through 191 which an enterprise can reach their workloads hosted in the Cloud 192 DCs. For example, AWS (Amazon Web Services) offers the following 193 options to reach workloads in AWS Cloud DCs: 195 - Internet gateway for any external entities to reach the 196 workloads hosted in AWS Cloud DC via the Internet. 197 - Virtual gateway (vGW) where IPsec tunnels [RFC6071] are 198 established between an enterprise's own gateway and AWS vGW, so 199 that the communications between those gateways can be secured 200 from the underlay (which might be the public Internet). 201 - Direct Connect, which allows enterprises to purchase direct 202 connect from network service providers to get a private leased 203 line interconnecting the enterprises gateway(s) and the AWS 204 Direct Connect routers. Via Direct Connect, an AWS Transit 205 Gateway can be used to interconnect multiple VPCs in different 206 Availability Zones. 208 CPEs at one Enterprise branch office are connected to the Internet 209 to reach AWS's vGW via IPsec tunnels. Other ports of such CPEs are 210 connected to AWS DirectConnect via a private network (without any 211 encryption). 212 +------------------------+ 213 | ,---. ,---. | 214 | (TN-1 ) ( TN-2)| 215 | `-+-' +--+ `-+-' | 216 | +----|vR|-----+ | 217 | ++-+ | 218 | | +-+----+ 219 | | /Internet\ For External 220 | +-------+ Gateway +---------------------- 221 | \ / to reach via Internet 222 | +-+----+ 223 | | 224 +------------------------+ 226 +------------------------+ 227 | ,---. ,---. | 228 | (TN-1 ) ( TN-2)| 229 | `-+-' +--+ `-+-' | 230 | +----|vR|-----+ | 231 | ++-+ | 232 | | +-+----+ 233 | | / virtual\ For IPsec Tunnel 234 | +-------+ Gateway +---------------------- 235 | \ / termination 236 | +-+----+ 237 | | 238 +------------------------+ 240 +------------------------+ 241 | ,---. ,---. | 242 | (TN-1 ) ( TN-2)| 243 | `-+-' +--+ `-+-' | 244 | +----|vR|-----+ | 245 | ++-+ | 246 | | +-+----+ +------+ 247 | | / \ For Direct /customer\ 248 | +-------+ Gateway +----------+ gateway | 249 | \ / Connect \ / 250 | +-+----+ +------+ 251 | | 252 +------------------------+ 254 Figure 1: Examples of Cloud DC connections. 256 3.2. Interconnect to Hybrid Cloud DCs 258 According to Gartner, by 2020 "hybrid will be the most common usage 259 of the cloud" as more enterprises see the benefits of integrating 260 public and private cloud infrastructures. However, enabling the 261 growth of hybrid cloud deployments in the enterprise requires fast 262 and safe interconnection between public and private cloud services. 263 For an enterprise to connect to applications & workloads hosted in 264 multiple Cloud DCs, the enterprise can use IPsec tunnels established 265 over the Internet or a (virtualized) leased line service to connect 266 its on-premises gateways to each of the Cloud DC's gateways, virtual 267 routers instantiated in the Cloud DCs, or any other suitable design 268 (including a combination thereof). 270 Some enterprises prefer to instantiate their own virtual 271 CPEs/routers inside the Cloud DC to connect the workloads within the 272 Cloud DC. Then an overlay path is established between customer 273 gateways to the virtual CPEs/routers for reaching the workloads 274 inside the cloud DC. 276 3.3. Connecting workloads among hybrid Cloud DCs 278 There are multiple approaches to interconnect workloads among 279 different Cloud DCs: 281 - Utilize Cloud DC provided transit gateways, which usually does 282 not work if Cloud DCs are owned and managed by different Cloud 283 providers. 284 - Hairpin all the traffic through the customer gateway, which 285 creates additional transmission delay & incurs cost when 286 exiting Cloud DCs, or 287 - Establish direct tunnels among different VPCs (Virtual Private 288 Clouds) via client's own virtual routers instantiated within 289 Cloud DCs. DMVPN (Dynamic Multipoint Virtual Private Network) 290 or DSVPN (Dynamic Smart VPN) techniques can be used to 291 establish direct Multi-point-to-Point or multi-point-to multi- 292 point tunnels among those client's own virtual routers. 294 DMVPN & DSVPN use NHRP (Next Hop Resolution Protocol) [RFC2735] so 295 that spoke nodes can register their IP addresses & WAN ports with 296 the hub node. The IETF ION (Internetworking over NBMA (non-broadcast 297 multiple access) WG standardized NHRP for connection-oriented NBMA 298 network (such as ATM) network address resolution more than two 299 decades ago. 301 There are many differences between virtual routers in Public Cloud 302 DCs and the nodes in an NBMA network. NHRP & DSVPN are not cannot be 303 used for registering virtual routers in Cloud DCs unless an 304 extension of such protocols is developed for that purpose. Other 305 protocols such as BGP can be used, as described in [BGP-SDWAN]. 307 4. Desired Properties for Networks that interconnect Hybrid Clouds 308 The networks that interconnect hybrid cloud DCs must address the 309 following requirements: 310 - High availability at any time, whatever the duration of the 311 connection to the cloud DC. 312 Many enterprises include cloud infrastructures in their 313 disaster recovery strategy, e.g., by enforcing periodic backup 314 policies within the cloud, or by running backup applications in 315 the Cloud, etc. Therefore, the connection to the cloud DCs may 316 not be permanent, but rather needs to be on-demand. 318 - Global reachability from different geographical zones, thereby 319 facilitating the proximity of applications as a function of the 320 end users' location, to improve latency. 321 - Elasticity and mobility, to instantiate additional applications 322 at Cloud DCs when end-users' usages increase and shut down 323 applications at locations when there are fewer end-users. 324 Some enterprises have front-end web portals running in cloud 325 DCs and database servers in their on-premises DCs. Those Front- 326 end web portals need to be reachable from the public Internet. 327 The backend connection to the sensitive data in database 328 servers hosted in the on-premises DCs might need secure 329 connections. 331 - Scalable security management. IPsec is commonly used to 332 interconnect cloud gateways with CPEs deployed in the 333 enterprise premises. For enterprises with a large number or 334 branch offices, managing the IPsec's Security Associations 335 among many nodes can be very difficult. 337 5. Problems with MPLS-based VPNs extending to Hybrid Cloud DCs 339 Traditional MPLS-based VPNs have been widely deployed as an 340 effective way to support businesses and organizations that require 341 network performance and reliability. MPLS shifted the burden of 342 managing a VPN service from enterprises to service providers. The 343 CPEs attached to MPLS VPNs are also simpler and less expensive, 344 since they do not need to manage routes to remote sites; they simply 345 pass all outbound traffic to the MPLS VPN PEs to which the CPEs are 346 attached (albeit multi-homing scenarios require more processing 347 logic on CPEs). MPLS has addressed the problems of scale, 348 availability, and fast recovery from network faults, and 349 incorporated traffic-engineering capabilities. 351 However, traditional MPLS-based VPN solutions are sub-optimized for 352 connecting end-users to dynamic workloads/applications in cloud DCs 353 because: 355 - The Provider Edge (PE) nodes of the enterprise's VPNs might not 356 have direct connections to third party cloud DCs that are used 357 for hosting workloads with the goal of providing an easy access 358 to enterprises' end-users. 360 - It usually takes some time to deploy provider edge (PE) routers 361 at new locations. When enterprise's workloads are changed from 362 one cloud DC to another (i.e., removed from one DC and re- 363 instantiated to another location when demand changes), the 364 enterprise branch offices need to be connected to the new cloud 365 DC, but the network service provider might not have PEs located 366 at the new location. 368 One of the main drivers for moving workloads into the cloud is 369 the widely available cloud DCs at geographically diverse 370 locations, where apps can be instantiated so that they can be 371 as close to their end-users as possible. When the user base 372 changes, the applications may be migrated to a new cloud DC 373 location closest to the new user base. 375 - Most of the cloud DCs do not expose their internal networks, so 376 the MPLS-based VPNs can only reach Cloud DC's Gateways, not to 377 the workloads hosted inside. 379 - Many cloud DCs use an overlay to connect their gateways to the 380 workloads located inside the DC. There has not been any 381 standard to address the interworking between the Cloud Overlay 382 and the enterprise' existing underlay networks. 384 Another roadblock is the lack of a standard way to express and 385 enforce consistent security policies for workloads that not only use 386 virtual addresses, but in which are also very likely hosted in 387 different locations within the Cloud DC [RFC8192]. The current VPN 388 path computation and bandwidth allocation schemes may not be 389 flexible enough to address the need for enterprises to rapidly 390 connect to dynamically instantiated (or removed) workloads and 391 applications regardless of their location/nature (i.e., third party 392 cloud DCs). 394 6. Problem with using IPsec tunnels to Cloud DCs 395 As described in the previous section, many Cloud operators expose 396 their gateways for external entities (which can be enterprises 397 themselves) to directly establish IPsec tunnels. Enterprises can 398 also instantiate virtual routers within Cloud DCs to connect to 399 their on-premises devices via IPsec tunnels. If there is only one 400 enterprise location that needs to reach the Cloud DC, an IPsec 401 tunnel is a very convenient solution. 403 However, many medium-to-large enterprises usually have multiple 404 sites and multiple data centers. For workloads and apps hosted in 405 cloud DCs, multiple sites need to communicate securely with those 406 cloud workloads and apps. This section documents some of the issues 407 associated with using IPsec tunnels to connect enterprise premises 408 with cloud gateways. 410 6.1. Complexity of multi-point any-to-any interconnection 412 The dynamic workload instantiated in cloud DC needs to communicate 413 with multiple branch offices and on-premises data centers. Most 414 enterprises need multi-point interconnection among multiple 415 locations, which can be provided by means of MPLS L2/L3 VPNs. 417 Using IPsec overlay paths to connect all branches & on-premises data 418 centers to cloud DCs requires CPEs to manage routing among Cloud DCs 419 gateways and the CPEs located at other branch locations, which can 420 dramatically increase the complexity of the design, possibly at the 421 cost of jeopardizing the CPE performance. 423 The complexity of requiring CPEs to maintain routing among other 424 CPEs is one of the reasons why enterprises migrated from Frame Relay 425 based services to MPLS-based VPN services. 427 MPLS-based VPNs have their PEs directly connected to the CPEs. 428 Therefore, CPEs only need to forward all traffic to the directly 429 attached PEs, which are therefore responsible for enforcing the 430 routing policy within the corresponding VPNs. Even for multi-homed 431 CPEs, the CPEs only need to forward traffic among the directly 432 connected PEs. However, when using IPsec tunnels between CPEs and 433 Cloud DCs, the CPEs need to compute, select, establish and maintain 434 routes for traffic to be forwarded to Cloud DCs, to remote CPEs via 435 VPN, or directly. 437 6.2. Poor performance over long distance 439 When enterprise CPEs or gateways are far away from cloud DC gateways 440 or across country/continent boundaries, performance of IPsec tunnels 441 over the public Internet can be problematic and unpredictable. Even 442 though there are many monitoring tools available to measure delay 443 and various performance characteristics of the network, the 444 measurement for paths over the Internet is passive and past 445 measurements may not represent future performance. 447 Many cloud providers can replicate workloads in different available 448 zones. An App instantiated in a cloud DC closest to clients may have 449 to cooperate with another App (or its mirror image) in another 450 region or database server(s) in the on-premises DC. This kind of 451 coordination requires predicable networking behavior/performance 452 among those locations. 454 6.3. Scaling Issues with IPsec Tunnels 456 IPsec can achieve secure overlay connections between two locations 457 over any underlay network, e.g., between CPEs and Cloud DC Gateways. 459 If there is only one enterprise location connected to the cloud 460 gateway, a small number of IPsec tunnels can be configured on-demand 461 between the on-premises DC and the Cloud DC, which is an easy and 462 flexible solution. 464 However, for multiple enterprise locations to reach workloads hosted 465 in cloud DCs, the cloud DC gateway needs to maintain multiple IPsec 466 tunnels to all those locations (e.g., as a hub & spoke topology). 467 For a company with hundreds or thousands of locations, there could 468 be hundreds (or even thousands) of IPsec tunnels terminating at the 469 cloud DC gateway, which is not only very expensive (because Cloud 470 Operators usually charge their customers based on connections), but 471 can be very processing intensive for the gateway. Many cloud 472 operators only allow a limited number of (IPsec) tunnels & bandwidth 473 to each customer. Alternatively, you could use a solution like 474 group encryption where a single IPsec SA is necessary at the GW but 475 the drawback here is key distribution and maintenance of a key 476 server, etc. 478 7. Problems of Using SD-WAN to connect to Cloud DCs 479 SD-WAN can establish parallel paths over multiple underlay networks 480 between two locations on-demand, for example, to support the 481 connections established between two CPEs interconnected by a 482 traditional MPLS VPN ([RFC4364] or [RFC4664]) or by IPsec [RFC6071] 483 tunnels. 485 SD-WAN lets enterprises augment their current VPN network with cost- 486 effective, readily available Broadband Internet connectivity, 487 enabling some traffic offloading to paths over the Internet 488 according to differentiated, possibly application-based traffic 489 forwarding policies, or when the MPLS VPN connection between the two 490 locations is congested, or otherwise undesirable or unavailable. 492 7.1. SD-WAN among branch offices vs. interconnect to Cloud DCs 494 SD-WAN interconnection of branch offices is not as simple as it 495 appears. For an enterprise with multiple sites, using SD-WAN overlay 496 paths among sites requires each CPE to manage all the addresses that 497 local hosts have the potential to reach, i.e., map internal VPN 498 addresses to appropriate SD-WAN paths. This is similar to the 499 complexity of Frame Relay based VPNs, where each CPE needed to 500 maintain mesh routing for all destinations if they were to avoid an 501 extra hop through a hub router. Even though SD-WAN CPEs can get 502 assistance from a central controller (instead of running a routing 503 protocol) to resolve the mapping between destinations and SD-WAN 504 paths, SD-WAN CPEs are still responsible for routing table 505 maintenance as remote destinations change their attachments, e.g., 506 the dynamic workload in other DCs are de-commissioned or added. 508 Even though originally envisioned for interconnecting branch 509 offices, SD-WAN offers a very attractive way for enterprises to 510 connect to Cloud DCs. 512 The SD-WAN for interconnecting branch offices and the SD-WAN for 513 interconnecting to Cloud DCs have some differences: 515 - SD-WAN for interconnecting branch offices usually have two end- 516 points (e.g., CPEs) controlled by one entity (e.g., a 517 controller or management system operated by the enterprise). 518 - SD-WAN for Cloud DC interconnects may consider CPEs owned or 519 managed by the enterprise, while remote end-points are being 520 managed or controlled by Cloud DCs (For the ease of 521 description, let's call such CPEs asymmetrically-managed CPEs). 523 - Cloud DCs may have different entry points (or devices) with one 524 entry point that terminates a private direct connection (based 525 upon a leased line for example) and other entry points being 526 devices terminating the IPsec tunnels, as shown in Figure 2. 528 Therefore, the SD-WAN design becomes asymmetric. 529 +------------------------+ 530 | ,---. ,---. | 531 | (TN-1 ) ( TN-2)| 532 | `-+-' +---+ `-+-' | 533 | +----|vR1|----+ | 534 | ++--+ | 535 | | +-+----+ 536 | | /Internet\ One path via 537 | +-------+ Gateway +---------------------+ 538 | \ / Internet \ 539 | +-+----+ \ 540 +------------------------+ \ 541 \ 542 +------------------------+ native traffic \ 543 | ,---. ,---. | without encryption| 544 | (TN-3 ) ( TN-4)| | 545 | `-+-' +--+ `-+-' | | +------+ 546 | +----|vR|-----+ | +------+ CPE | 547 | ++-+ | | +------+ 548 | | +-+----+ | 549 | | / virtual\ One path via IPsec Tunnel | 550 | +-------+ Gateway +-------------------------- + 551 | \ / Encrypted traffic over| 552 | +-+----+ public network | 553 +------------------------+ | 554 | 555 +------------------------+ | 556 | ,---. ,---. | Native traffic | 557 | (TN-5 ) ( TN-6)| without encryption | 558 | `-+-' +--+ `-+-' | over secure network| 559 | +----|vR|-----+ | | 560 | ++-+ | | 561 | | +-+----+ +------+ | 562 | | / \ Via Direct /customer\ | 563 | +-------+ Gateway +----------+ gateway |-----+ 564 | \ / Connect \ / 565 | +-+----+ +------+ 566 +------------------------+ 568 Figure 2: Different Underlays to Reach Cloud DC 570 8. End-to-End Security Concerns for Data Flows 572 When IPsec tunnels established from enterprise on-premises CPEs 573 are terminated at the Cloud DC gateway where the workloads or 574 applications are hosted, some enterprises have concerns regarding 575 traffic to/from their workload being exposed to others behind the 576 data center gateway (e.g., exposed to other organizations that 577 have workloads in the same data center). 578 To ensure that traffic to/from workloads is not exposed to 579 unwanted entities, IPsec tunnels may go all the way to the 580 workload (servers, or VMs) within the DC. 582 9. Requirements for Dynamic Cloud Data Center VPNs 584 In order to address the aforementioned issues, any solution for 585 enterprise VPNs that includes connectivity to dynamic workloads or 586 applications in cloud data centers should satisfy a set of 587 requirements: 589 - The solution should allow enterprises to take advantage of the 590 current state-of-the-art in VPN technology, in both traditional 591 MPLS-based VPNs and IPsec-based VPNs (or any combination 592 thereof) that run over the public Internet. 593 - The solution should not require an enterprise to upgrade all 594 their existing CPEs. 595 - The solution should support scalable IPsec key management among 596 all nodes involved in DC interconnect schemes. 597 - The solution needs to support easy and fast, on-the-fly, VPN 598 connections to dynamic workloads and applications in third 599 party data centers, and easily allow these workloads to migrate 600 both within a data center and between data centers. 601 - Allow VPNs to provide bandwidth and other performance 602 guarantees. 603 - Be a cost-effective solution for enterprises to incorporate 604 dynamic cloud-based applications and workloads into their 605 existing VPN environment. 607 10. Security Considerations 609 The draft discusses security requirements as a part of the problem 610 space, particularly in sections 4, 5, and 8. 612 Solution drafts resulting from this work will address security 613 concerns inherent to the solution(s), including both protocol 614 aspects and the importance (for example) of securing workloads in 615 cloud DCs and the use of secure interconnection mechanisms. 617 IANA Considerations 619 This document requires no IANA actions. RFC Editor: Please remove 620 this section before publication. 622 11. References 624 11.1. Normative References 626 11.2. Informative References 628 [RFC2735] B. Fox, et al "NHRP Support for Virtual Private 629 networks". Dec. 1999. 631 [RFC8192] S. Hares, et al "Interface to Network Security Functions 632 (I2NSF) Problem Statement and Use Cases", July 2017 634 [ITU-T-X1036] ITU-T Recommendation X.1036, "Framework for creation, 635 storage, distribution and enforcement of policies for 636 network security", Nov 2007. 638 [RFC6071] S. Frankel and S. Krishnan, "IP Security (IPsec) and 639 Internet Key Exchange (IKE) Document Roadmap", Feb 2011. 641 [RFC4364] E. Rosen and Y. Rekhter, "BGP/MPLS IP Virtual Private 642 Networks (VPNs)", Feb 2006 644 [RFC4664] L. Andersson and E. Rosen, "Framework for Layer 2 Virtual 645 Private Networks (L2VPNs)", Sept 2006. 647 [BGP-SDWAN] L. Dunbar, et al. "BGP Extension for SDWAN Overlay 648 Networks", draft-dunbar-idr-bgp-sdwan-overlay-ext-03, 649 work-in-progress, Nov 2018. 651 12. Acknowledgments 653 Many thanks to Ignas Bagdonas, Michael Huang, Liu Yuan Jiao, 654 Katherine Zhao, and Jim Guichard for the discussion and 655 contributions. 657 Authors' Addresses 659 Linda Dunbar 660 Huawei 661 Email: Linda.Dunbar@huawei.com 663 Andrew G. Malis 664 Huawei 665 Email: agmalis@gmail.com 667 Christian Jacquenet 668 Orange 669 Rennes, 35000 670 France 671 Email: Christian.jacquenet@orange.com 673 Mehmet Toy 674 Verizon 675 One Verizon Way 676 Basking Ridge, NJ 07920 677 Email: mehmet.toy@verizon.com