idnits 2.17.1 draft-dong-teas-enhanced-vpn-02.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 : ---------------------------------------------------------------------------- No issues found here. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (October 18, 2018) is 2017 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- == Unused Reference: 'RFC2119' is defined on line 1058, but no explicit reference was found in the text == Outdated reference: A later version (-13) exists of draft-ietf-detnet-architecture-08 == Outdated reference: A later version (-02) exists of draft-ietf-detnet-dp-sol-ip-00 == Outdated reference: A later version (-20) exists of draft-ietf-detnet-use-cases-19 == Outdated reference: A later version (-24) exists of draft-ietf-teas-actn-vn-yang-02 == Outdated reference: A later version (-13) exists of draft-lee-teas-te-service-mapping-yang-12 Summary: 0 errors (**), 0 flaws (~~), 7 warnings (==), 1 comment (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 TEAS Working Group J. Dong 3 Internet-Draft S. Bryant 4 Intended status: Informational Huawei 5 Expires: April 21, 2019 Z. Li 6 China Mobile 7 T. Miyasaka 8 KDDI Corporation 9 October 18, 2018 11 A Framework for Enhanced Virtual Private Networks (VPN+) 12 draft-dong-teas-enhanced-vpn-02 14 Abstract 16 This document specifies a framework for using existing, modified and 17 potential new networking technologies as components to provide an 18 enhanced VPN (VPN+) service. The purpose is to enable virtual 19 private networks (VPNs) to support the needs of new applications, 20 particularly applications that are associated with 5G services. A 21 network enhanced with these properties can form the underpinning of 22 network slicing, but will also be of use in its own right. 24 Status of This Memo 26 This Internet-Draft is submitted in full conformance with the 27 provisions of BCP 78 and BCP 79. 29 Internet-Drafts are working documents of the Internet Engineering 30 Task Force (IETF). Note that other groups may also distribute 31 working documents as Internet-Drafts. The list of current Internet- 32 Drafts is at https://datatracker.ietf.org/drafts/current/. 34 Internet-Drafts are draft documents valid for a maximum of six months 35 and may be updated, replaced, or obsoleted by other documents at any 36 time. It is inappropriate to use Internet-Drafts as reference 37 material or to cite them other than as "work in progress." 39 This Internet-Draft will expire on April 21, 2019. 41 Copyright Notice 43 Copyright (c) 2018 IETF Trust and the persons identified as the 44 document authors. All rights reserved. 46 This document is subject to BCP 78 and the IETF Trust's Legal 47 Provisions Relating to IETF Documents 48 (https://trustee.ietf.org/license-info) in effect on the date of 49 publication of this document. Please review these documents 50 carefully, as they describe your rights and restrictions with respect 51 to this document. Code Components extracted from this document must 52 include Simplified BSD License text as described in Section 4.e of 53 the Trust Legal Provisions and are provided without warranty as 54 described in the Simplified BSD License. 56 Table of Contents 58 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 3 59 2. Overview of the Requirements . . . . . . . . . . . . . . . . 4 60 2.1. Isolation between Virtual Networks . . . . . . . . . . . 4 61 2.1.1. A Pragmatic Approach to Isolation . . . . . . . . . . 6 62 2.2. Performance Guarantee . . . . . . . . . . . . . . . . . . 7 63 2.3. Integration . . . . . . . . . . . . . . . . . . . . . . . 8 64 2.4. Dynamic Configuration . . . . . . . . . . . . . . . . . . 9 65 2.5. Customized Control . . . . . . . . . . . . . . . . . . . 9 66 2.6. Applicability . . . . . . . . . . . . . . . . . . . . . . 10 67 3. Architecture of Enhanced VPN . . . . . . . . . . . . . . . . 10 68 3.1. Layered Architecture . . . . . . . . . . . . . . . . . . 11 69 3.2. Multi-Point to Multi-Point . . . . . . . . . . . . . . . 13 70 3.3. Application Specific Network Types . . . . . . . . . . . 13 71 4. Candidate Technologies . . . . . . . . . . . . . . . . . . . 13 72 4.1. Underlay Data Plane . . . . . . . . . . . . . . . . . . . 14 73 4.1.1. FlexE . . . . . . . . . . . . . . . . . . . . . . . . 14 74 4.1.2. Dedicated Queues . . . . . . . . . . . . . . . . . . 15 75 4.1.3. Time Sensitive Networking . . . . . . . . . . . . . . 15 76 4.2. Network Layer . . . . . . . . . . . . . . . . . . . . . . 16 77 4.2.1. Deterministic Networking . . . . . . . . . . . . . . 16 78 4.2.2. MPLS Traffic Engineering (MPLS-TE) . . . . . . . . . 16 79 4.2.3. Segment Routing . . . . . . . . . . . . . . . . . . . 16 80 4.3. Control Plane . . . . . . . . . . . . . . . . . . . . . . 18 81 4.4. Management Plane . . . . . . . . . . . . . . . . . . . . 19 82 5. Scalability Considerations . . . . . . . . . . . . . . . . . 19 83 5.1. Maximum Stack Depth of SR . . . . . . . . . . . . . . . . 20 84 5.2. RSVP Scalability . . . . . . . . . . . . . . . . . . . . 21 85 6. OAM Considerations . . . . . . . . . . . . . . . . . . . . . 21 86 7. Enhanced Resiliency . . . . . . . . . . . . . . . . . . . . . 21 87 8. Security Considerations . . . . . . . . . . . . . . . . . . . 23 88 9. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 23 89 10. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 23 90 11. References . . . . . . . . . . . . . . . . . . . . . . . . . 23 91 11.1. Normative References . . . . . . . . . . . . . . . . . . 23 92 11.2. Informative References . . . . . . . . . . . . . . . . . 23 93 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 27 95 1. Introduction 97 Virtual private networks (VPNs) have served the industry well as a 98 means of providing different groups of users with logically isolated 99 access to a common network. The common or base network that is used 100 to provide the VPNs is often referred to as the underlay, and the VPN 101 is often called an overlay. 103 Customers of a network operator may request enhanced VPN services 104 with additional characteristics such as complete isolation from other 105 VPNs so that changes in network load have no effect on the throughput 106 or latency of the service provided to the customer. 108 Driven largely by needs surfacing from 5G, the concept of network 109 slicing has gained traction.[NGMN-NS-Concept] [TS23501] [TS28530] 110 [BBF-SD406]. Network slicing requires the transport network to 111 support partitioning the network resources to provide the client with 112 dedicated (private) networking, computing, and storage resources 113 drawn from a shared pool. The slices may be seen as (and operated 114 as) virtual networks. 116 Thus, there is a need to create virtual networks with enhanced 117 characteristics. The tenant of such a virtual network can require a 118 degree of isolation and performance that previously could only be 119 satisfied by dedicated networks. Additionally the tenant may ask for 120 some level of control to their virtual network e.g. to customize the 121 service paths in the network slice. 123 These properties cannot be met with pure overlay networks, as they 124 require tighter coordination and integration between the underlay and 125 the overlay network. This document introduces a new network service 126 called enhanced VPN (VPN+). VPN+ refers to a virtual network which 127 has dedicated network resources allocated from the underlay network. 128 Unlike a traditional VPN, an enhanced VPN can achieve greater 129 isolation and guaranteed performance. These new properties, which 130 have general applicability, may also be of interest as part of a 131 network slicing solution. 133 This document specifies a framework for using existing, modified and 134 potential new networking technologies as components to provide an 135 enhanced VPN (VPN+) service. Specifically we are concerned with: 137 o The design of the enhanced data plane 139 o The necessary protocols in both underlay and the overlay of 140 enhanced VPN 142 o The mechanisms to achieve integration between overlay and underlay 143 o The necessary Operation, Administration and Management (OAM) 144 methods to instrument an enhanced VPN to make sure that the 145 required Service Level Agreement (SLA) are met, and to take any 146 required action to avoid SLA violation, such as switching to an 147 alternate path 149 The required network layered structure to achieve this is shown in 150 Section 3.1. 152 One use for enhanced VPNs is to create network slices with different 153 isolation requirements. Such network slices may be used to provide 154 different tenants of vertical industrial markets with their own 155 virtual network with the explicit characteristics required. These 156 network slices may be "hard" slices providing a high degree of 157 confidence that the VPN+ characteristics will be maintained over the 158 slice life cycle, or they may be "soft" slices in which case some 159 interaction may be experienced. 161 2. Overview of the Requirements 163 In this section we provide an overview of the requirements of an 164 enhanced VPN. 166 2.1. Isolation between Virtual Networks 168 Isolation is a feature of the services offered by a network operator 169 where the traffic from one service instance is isolated from the 170 traffic of other services. There are different grades of isolation 171 that range from simple separation of traffic on delivery (ensuring 172 that traffic is not delivered to the wrong customer) all the way to 173 complete separation within the underlay so that the traffic from 174 different services use distinct network resources. 176 We introduce the terms hard and soft isolation to cover some of these 177 cases. A VPN has soft isolation if the traffic of one VPN cannot be 178 inspected by the traffic of another. Both IP and MPLS VPNs are 179 examples of soft isolated VPNs because the network delivers the 180 traffic only to the required VPN endpoints. However, the traffic 181 from one or more VPNs and regular network traffic may congest the 182 network resulting in packet loss and delay for other VPNs operating 183 normally. The ability for a VPN to be sheltered from this effect is 184 called hard isolation, and this property is required by some critical 185 applications. Although discussion of these isolation requirements is 186 triggered by the needs of 5G networks, they have general utility. 188 The requirement is for an operator to provide both hard and soft 189 isolation between the tenants/applications using one enhanced VPN and 190 the tenants/applications using another enhanced VPN. Hard isolation 191 is needed so that applications with exacting requirements can 192 function correctly, despite other demands (perhaps a burst on another 193 VPN) competing for the underlying resources. In practice isolation 194 may be offered as a spectrum between soft and hard. 196 An example of hard isolation is a network supporting both emergency 197 services and public broadband multi-media services. During a major 198 incident the VPNs supporting these services would both be expected to 199 experience high data volumes, and it is important that both make 200 progress in the transmission of their data. In these circumstances 201 the VPNs would require an appropriate degree of isolation to be able 202 to continue to operate acceptably. 204 In order to provide the required isolation, resources may have to be 205 reserved in the data plane of the underlay network and dedicated to 206 traffic from a specific VPN. This may introduce scalability 207 concerns, thus some trade-off needs to be considered to provide the 208 required isolation between network slices while still allowing 209 reasonable sharing inside each network slice. 211 An optical layer can offer a high degree of isolation, at the cost of 212 allocating resources on a long term and end-to-end basis. Such an 213 arrangement means that the full cost of the resources must be borne 214 by the service that is allocated with the resources. On the other 215 hand, where adequate isolation can be achieved at the packet layer, 216 this permits the resources to be shared amongst many services and 217 only dedicated to a service on a temporary basis. This in turn, 218 allows greater statistical multiplexing of network resources and thus 219 amortizes the cost over many services, leading to better economy. 220 However, the degree of isolation required by network slicing cannot 221 be entirely met with existing mechanisms such as Traffic Engineered 222 Label Switched Paths (TE-LSPs). This is because most implementations 223 enforce the bandwidth in the data-plane only at the PEs, but at the P 224 routers the bandwidth is only reserved in the control plane, thus 225 bursts of data can accidentally occur at a P router with higher than 226 committed data rate. 228 There are several new technologies that provide some assistance with 229 these data plane issues. Firstly there is the IEEE project on Time 230 Sensitive Networking [TSN] which introduces the concept of packet 231 scheduling of delay and loss sensitive packets. Then there is 232 [FLEXE] which provides the ability to multiplex multiple channels 233 over one or more Ethernet links in a way that provides hard 234 isolation. Finally there are advanced queueing approaches which 235 allow the construction of virtual sub-interfaces, each of which is 236 provided with dedicated resource in a shared physical interface. 237 These approaches are described in more detail later in this document. 239 In the remainder of this section we explore how isolation may be 240 achieved in packet networks. 242 2.1.1. A Pragmatic Approach to Isolation 244 A key question is whether it is possible to achieve hard isolation in 245 packet networks, which were never designed to support hard isolation. 246 On the contrary, they were designed to provide statistical 247 multiplexing, a significant economic advantage when compared to a 248 dedicated, or a Time Division Multiplexing (TDM) network. However 249 there is no need to provide any harder isolation than is required by 250 the application. Pseudowires [RFC3985] emulate services that would 251 have had hard isolation in their native form. An approximation to 252 this requirement is sufficient in most cases. 254 Thus, for example, using FlexE or a channelized sub-interface 255 together with packet scheduling as interface slicing, optionally 256 along with the slicing of node resources, a type of hard isolation 257 can be provided that is adequate for many VPN+ applications. Other 258 applications may be either satisfied with a classical VPN with or 259 without reserved bandwidth, or may need dedicated point to point 260 fiber. The needs of each application must be quantified in order to 261 provide an economic solution that satisfies those needs without over- 262 engineering. 264 This spectrum of isolation is shown below: 266 O=================================================O 267 | \---------------v---------------/ 268 Statistical Pragmatic Absolute 269 Multiplexing Isolation Isolation 270 (Traditional VPNs) (Enhanced VPN) (Dedicated Network) 272 At one end of the above figure, we have traditional statistical 273 multiplexing technologies that support VPNs. This is a service type 274 that has served the industry well and will continue to do so. At the 275 opposite end of the spectrum we have the absolute isolation provided 276 by traditional transport networks. The goal of enhanced VPN is 277 pragmatic isolation. This is isolation that is better than is 278 obtainable from pure statistical multiplexing, more cost effective 279 and flexible than a dedicated network, but which is a practical 280 solution that is good enough for the majority of applications. 282 2.2. Performance Guarantee 284 There are several kinds of performance guarantees, including 285 guaranteed maximum packet loss, guaranteed maximum delay and 286 guaranteed delay variation. 288 Guaranteed maximum packet loss is a common parameter, and is usually 289 addressed by setting the packet priorities, queue size and discard 290 policy. However this becomes more difficult when the requirement is 291 combined with the latency requirement. The limiting case is zero 292 congestion loss, and that is the goal of the Deterministic Networking 293 work that the IETF [DETNET] and IEEE [TSN] are pursuing. In modern 294 optical networks, loss due to transmission errors is already 295 approaches zero, but there are the possibilities of failure of the 296 interface or the fiber itself. This can only be addressed by some 297 form of signal duplication and transmission over diverse paths. 299 Guaranteed maximum latency is required in a number of applications 300 particularly real-time control applications and some types of virtual 301 reality applications. The work of the IETF Deterministic Networking 302 (DetNet) Working Group [DETNET] is relevant; however the scope needs 303 to be extended to methods of enhancing the underlay to better support 304 the delay guarantee, and to integrate these enhancements with the 305 overall service provision. 307 Guaranteed maximum delay variation is a service that may also be 308 needed. [I-D.ietf-detnet-use-cases] calls up a number of cases where 309 this is needed, for example electrical utilities have an operational 310 need for this. Time transfer is one example of a service that needs 311 this, although it is in the nature of time that the service might be 312 delivered by the underlay as a shared service and not provided 313 through different virtual networks. Alternatively a dedicated 314 virtual network may be used to provide this as a shared service. 316 This suggests that a spectrum of service guarantee be considered when 317 deploying an enhanced VPN. As a guide to understanding the design 318 requirements we can consider four types: 320 o Best effort 322 o Assured bandwidth 324 o Guaranteed latency 326 o Enhanced delivery 328 Best effort service is the basic service that current VPNs can 329 provide. 331 An assured bandwidth service is one in which the bandwidth over some 332 period of time is assured, this can be achieved either simply based 333 on best effort with over-capacity provisioning, or it can be based on 334 TE-LSPs with bandwidth reservation. The instantaneous bandwidth is 335 however, not necessarily assured, depending on the technique used. 336 Providing assured bandwidth to VPNs, for example by using TE-LSPs, is 337 not widely deployed at least partially due to scalability concerns. 338 Guaranteed latency and enhanced delivery are not yet integrated with 339 VPNs. 341 A guaranteed latency service has a latency upper bound provided by 342 the network. Assuring the upper bound is more important than 343 achieving the minimum latency. 345 In Section 2.1 we considered the work of the IEEE Time Sensitive 346 Networking (TSN) project [TSN] and the work of the IETF DetNet 347 Working group [DETNET] in the context of isolation. The TSN and 348 DetNet work is of greater relevance in assuring end-to-end packet 349 latency. It is also of importance in considering enhanced delivery. 351 An enhanced delivery service is one in which the underlay network (at 352 layer 3) attempts to deliver the packet through multiple paths in the 353 hope of eliminating packet loss due to equipment or media failures. 355 It is these last two characteristics that an enhanced VPN adds to a 356 VPN service. 358 Flex Ethernet [FLEXE] is a useful underlay to provide these 359 guarantees. This is a method of providing time-slot based 360 channelization over an Ethernet bearer. Such channels are fully 361 isolated from other channels running over the same Ethernet bearer. 362 As noted elsewhere this produces hard isolation but makes the 363 reclamation of unused bandwidth more difficult. 365 These approaches can be used in tandem. It is possible to use FlexE 366 to provide tenant isolation, and then to use the TSN/Detnet approach 367 to provide a performance guarantee inside the a slice or tenant VPN. 369 2.3. Integration 371 A solution to the enhanced VPN problem has to provide close 372 integration of both overlay VPN and the underlay network resource. 373 This needs be done in a flexible and scalable way so that it can be 374 widely deployed in operator networks to support a reasonable number 375 of enhanced VPN customers. 377 Taking mobile networks and in particular 5G into consideration, the 378 integration of network and the service functions is a likely 379 requirement. The work in IETF SFC working group [SFC] provides a 380 foundation for this integration. 382 2.4. Dynamic Configuration 384 Enhanced VPNs need to be created, modified, and removed from the 385 network according to service demand. An enhanced VPN that requires 386 hard isolation must not be disrupted by the instantiation or 387 modification of another enhanced VPN. Determining whether 388 modification of an enhanced VPN can be disruptive to that VPN, and in 389 particular the traffic in flight will be disrupted can be a difficult 390 problem. 392 The data plane aspects of this problem are discussed further in 393 Section 4. 395 The control plane aspects of this problem are discussed further in 396 Section 4.3. 398 The management plane aspects of this problem are discussed further in 399 Section 4.4 401 Dynamic changes both to the VPN and to the underlay transport network 402 need to be managed to avoid disruption to sensitive services. 404 In addition to non-disruptively managing the network as a result of 405 gross change such as the inclusion of a new VPN endpoint or a change 406 to a link, VPN traffic might need to be moved as a result of traffic 407 volume changes. 409 2.5. Customized Control 411 In some cases it is desirable that an enhanced VPN has a customized 412 control plane, so that the tenant of the enhanced VPN can have some 413 control to the resources and functions allocated to this enhanced 414 VPN. For example, the tenant may be able to specify the service 415 paths in his own enhanced VPN. Depending on the requirement, an 416 enhanced VPN may have its own dedicated controller, or it may be 417 provided with an interface to a control system which is shared with a 418 set of other tenants, or it may be provided with an interface to the 419 control system provided by the network operator. 421 Further detail on this requirement will be provided in a future 422 version of the draft. 424 2.6. Applicability 426 The technologies described in this document should be applicable to a 427 number types of VPN services such as: 429 o Layer 2 point to point services such as pseudowires [RFC3985] 431 o Layer 2 VPNs [RFC4664] 433 o Ethernet VPNs [RFC7209] 435 o Layer 3 VPNs [RFC4364], [RFC2764] 437 Where such VPN types need enhanced isolation and delivery 438 characteristics the technology described here can be used to provide 439 an underlay with the required enhanced performance. 441 3. Architecture of Enhanced VPN 443 A number of enhanced VPN services will typically be provided by a 444 common network infrastructure. Each enhanced VPN consists of both 445 the overlay and a specific set of dedicated network resources and 446 functions allocated in the underlay to satisfy the needs of the VPN 447 tenant. The integration between overlay and various underlay 448 resources ensures the isolation between different enhanced VPNs, and 449 achieves the guaranteed performance for different services. 451 An enhanced VPN needs to be designed with consideration given to: 453 o A enhanced data plane 455 o A control plane to create enhanced VPN, making use of the data 456 plane isolation and guarantee techniques 458 o A management plane for enhanced VPN service life-cycle management 460 These required characteristics are expanded below: 462 o Enhanced data plane 464 * Provides the required resource isolation capability, e.g. 465 bandwidth guarantee. 467 * Provides the required packet latency and jitter characteristics 469 * Provides the required packet loss characteristics 470 * Provides the mechanism to identify network slice and the 471 associated resources 473 o Control plane 475 * Collect the underlying network topology and resources available 476 and export this to other nodes and/or the centralized 477 controller as required. 479 * Create the required virtual networks with the resource and 480 properties needed by the enhanced VPN services that are 481 assigned to it. 483 * Determine the risk of SLA violation and take appropriate 484 avoiding action 486 * Determine the right balance of per-packet and per-node state 487 according to the needs of enhanced VPN service to scale to the 488 required size 490 o Management plane 492 * Provides the life-cycle management (creation, modification, 493 decommissioning) of enhanced VPN 495 * Provide a interface between the enhanced VPN provider and the 496 enhanced VPN clients such that some of the operation requests 497 can be met without interfering other enhanced VPN clients. 499 This document will focus on the data plane and control plane of the 500 enhanced VPN. The details of the management plane is outside the 501 scope of this document. 503 3.1. Layered Architecture 505 The layered architecture of enhanced VPN is shown in Figure 1. 507 +-------------------+ } 508 | Network Controller| } Centralized 509 +-------------------+ } Control 510 . . . . . 511 . . . . . 512 . N----N----N . } 513 . / / . } 514 N-----N-----N----N-----N } 515 N----N } 516 / / \ } Virtual 517 N-----N----N----N-----N } Networks 518 N----N } 519 / / } 520 N-----N-----N----N-----N } 522 +----+ ===== +----+ ===== +----+ ===== +----+ } 523 +----+ ===== +----+ ===== +----+ ===== +----+ } Physical 524 +----+ ===== +----+ ===== +----+ ===== +----+ } Network 525 +----+ +----+ +----+ +----+ } 526 N L N L N L N 528 N = Partitioned node 529 L = Partitioned link 531 +----+ = Partition within a node 532 +----+ 534 ====== = Partition within a link 536 Figure 1: The Layered Architecture 538 Underpinning everything is the physical infrastructure layer 539 consisting of partitioned links and nodes which provide the 540 underlying resources used to provision the separated virtual 541 networks. Various components and techniques as discussed in 542 Section 4 can be used to provide the resource partition, such as 543 FlexE, Time Sensitive Networking, Deterministic Networking, etc. 544 These partitions may be physical, or virtual so long as the SLA 545 required by the higher layers is met. 547 These techniques can be used to provision the virtual networks with 548 dedicated resources that they need. To get the required 549 functionality there needs to be integration between these overlays 550 and the underlay providing the physical resources. 552 The centralized controller is used to create the virtual networks, to 553 allocate the resources to each virtual network and to provision the 554 enhanced VPN services within the virtual networks. A distributed 555 control plane may also be used for the distribution of the topology 556 and attribute information of the virtual networks. 558 The creation and allocation process needs to take a holistic view of 559 the needs of all of its tenants, and to partition the resources 560 accordingly. However within a virtual network these resources can if 561 required be managed via a dynamic control plane. This provides the 562 required scalability and isolation. 564 3.2. Multi-Point to Multi-Point 566 At the VPN service level, the connectivity are usually mesh or 567 partial-mesh. To support such kind of VPN service, the corresponding 568 underlay is also an abstract MP2MP medium. However when service 569 guarantees are provided, the point-to-point path through the underlay 570 of the enhanced VPN needs to be specifically engineered to meet the 571 required performance guarantees. 573 3.3. Application Specific Network Types 575 Although a lot of the traffic that will be carried over the enhanced 576 VPN will likely be IPv4 or IPv6, the design has to be capable of 577 carrying other traffic types, in particular Ethernet traffic. This 578 is easily accomplished through the various pseudowire (PW) techniques 579 [RFC3985]. Where the underlay is MPLS, Ethernet can be carried over 580 the enhanced VPN encapsulated according to the method specified in 581 [RFC4448]. Where the underlay is IP, Layer Two Tunneling Protocol - 582 Version 3 (L2TPv3) [RFC3931] can be used with Ethernet traffic 583 carried according to [RFC4719]. Encapsulations have been defined for 584 most of the common layer two type for both PW over MPLS and for 585 L2TPv3. 587 4. Candidate Technologies 589 A VPN is a network created by applying a multiplexing technique to 590 the underlying network (the underlay) in order to distinguish the 591 traffic of one VPN from that of another. A VPN path that travels by 592 other than the shortest path through the underlay normally requires 593 state in the underlay to specify that path. State is normally 594 applied to the underlay through the use of the RSVP Signaling 595 protocol, or directly through the use of an SDN controller, although 596 other techniques may emerge as this problem is studied. This state 597 gets harder to manage as the number of VPN paths increases. 598 Furthermore, as we increase the coupling between the underlay and the 599 overlay to support the enhanced VPN service, this state will increase 600 further. 602 In an enhanced VPN different subsets of the underlay resources are 603 dedicated to different enhanced VPNs. Any enhanced VPN solution thus 604 needs tighter coupling with underlay than is the case with existing 605 VPNs. We cannot for example share the tunnel between enhanced VPNs 606 which require hard isolation. 608 A number of candidate underlay data plane solutions which can be used 609 provide the required isolation and guarantee are described in 610 following sections. 612 o FlexE 614 o Time Sensitive Networking 616 o Dedicated Queues 618 We then consider the problem of slice differentiation and resource 619 representation in the network layer. The candidate technologies are: 621 o MPLS 623 o MPLS-SR 625 o Segment Routing over IPv6 (SRv6) 627 o Deterministic Networking 629 The considerations about the control plane is also described. 631 4.1. Underlay Data Plane 633 4.1.1. FlexE 635 FlexE [FLEXE] is a method of creating a point-to-point Ethernet with 636 a specific fixed bandwidth. FlexE provides the ability to multiplex 637 multiple channels over an Ethernet link in a way that provides hard 638 isolation. FlexE also supports the bonding of multiple links, which 639 can be used to create larger links out of multiple slower links in a 640 more efficient way that traditional link aggregation. FlexE also 641 supports the sub-rating of links, which allows an operator to only 642 use a portion of a link. However it is a only a link level 643 technology. When packets are received by the downstream node, they 644 need to be processed in a way that preserves that isolation in the 645 downstream node. This in turn requires a queuing and forwarding 646 implementation that preserves the end-to-end isolation. 648 If different FlexE channels are used for different services, then no 649 sharing is possible between the FlexE channels. This in turn means 650 that it may be difficult to dynamically redistribute unused bandwidth 651 to lower priority services. This may increase the cost of providing 652 services on the network. On the other hand, FlexE can be used to 653 provide hard isolation between different tenants on a shared 654 interface. The tenant can then use other methods to manage the 655 relative priority of their own traffic in each FlexE channel. 657 Methods of dynamically re-sizing FlexE channels and the implication 658 for enhanced VPN is for further study. 660 4.1.2. Dedicated Queues 662 In order to provide multiple isolated virtual networks for enhanced 663 VPN, the conventional Diff-Serv based queuing system [RFC2475] 664 [RFC4594] is insufficient, due to the limited number of queues which 665 cannot differentiate between traffic of different enhanced VPNs, and 666 the range of service classes that each need to provide to their 667 tenants. This problem is particularly acute with an MPLS underlay 668 due to the small number of traffic class services available. In 669 order to address this problem and reduce the interference between 670 enhanced VPNs, it is necessary to steer traffic of VPNs to dedicated 671 input and output queues. Routers usually have large amount of queues 672 and sophisticated queuing systems, which could be used or enhanced to 673 provide the levels of isolation required by the applications of 674 enhanced VPN. For example, on one physical interface, the queuing 675 system can provide a set of virtual sub-interfaces, each allocated 676 with dedicated queueing and buffer resources. Sophisticated queuing 677 systems of this type may be used to provide end-to-end virtual 678 isolation between traffic of different enhanced VPNs. 680 4.1.3. Time Sensitive Networking 682 Time Sensitive Networking (TSN) [TSN] is an IEEE project that is 683 designing a method of carrying time sensitive information over 684 Ethernet. It introduces the concept of packet scheduling where a 685 high priority packet stream may be given a scheduled time slot 686 thereby guaranteeing that it experiences no queuing delay and hence a 687 reduced latency. However, when no scheduled packet arrives, its 688 reserved time-slot is handed over to best effort traffic, thereby 689 improving the economics of the network. The mechanisms defined in 690 TSN can be used to meet the requirements of time sensitive services 691 of an enhanced VPN. 693 Ethernet can be emulated over a Layer 3 network using a pseudowire. 694 However the TSN payload would be opaque to the underlay and thus not 695 treated specifically as time sensitive data. The preferred method of 696 carrying TSN over a layer 3 network is through the use of 697 deterministic networking as explained in the following section of 698 this document. 700 4.2. Network Layer 702 4.2.1. Deterministic Networking 704 Deterministic Networking (DetNet) [I-D.ietf-detnet-architecture] is a 705 technique being developed in the IETF to enhance the ability of layer 706 3 networks to deliver packets more reliably and with greater control 707 over the delay. The design cannot use re-transmission techniques 708 such as TCP since that can exceed the delay tolerated by the 709 applications. Even the delay improvements that are achieved with 710 Stream Control Transmission Protocol Partial Reliability Extenstion 711 (SCTP-PR) [RFC3758] do not meet the bounds set by application 712 demands. DetNet pre-emptively sends copies of the packet over 713 various paths to minimize the chance of all packets being lost, and 714 trims duplicate packets to prevent excessive flooding of the network 715 and to prevent multiple packets being delivered to the destination. 716 It also seeks to set an upper bound on latency. The goal is not to 717 minimize latency; the optimum upper bound paths may not be the 718 minimum latency paths. 720 DetNet is based on flows. It currently does not specify the use of 721 underlay topology other than the base topology. To be of use for 722 enhanced VPN, DetNet needs to be integrated with different virtual 723 topologies of enhanced VPNs. 725 The detailed design that allows the use DetNet in a multi-tenant 726 network, and how to improve the scalability of DetNet in a multi- 727 tenant network are topics for further study. 729 4.2.2. MPLS Traffic Engineering (MPLS-TE) 731 MPLS-TE introduces the concept of reserving end-to-end bandwidth for 732 a TE-LSP, which can be used as the underlay of VPNs. It also 733 introduces the concept of non-shortest path routing through the use 734 of the Explicit Route Object [RFC3209]. VPN traffic can be run over 735 dedicated TE-LSPs to provide reserved bandwidth for each specific 736 connection in a VPN. Some network operators have concerns about the 737 scalability and management overhead of RSVP-TE system, and this has 738 lead them to consider other solutions for their networks. 740 4.2.3. Segment Routing 742 Segment Routing [RFC8402] is a method that prepends instructions to 743 packets at the head-end node and optionally at various points as it 744 passes though the network. These instructions allow the packets to 745 be routed on paths other than the shortest path for various traffic 746 engineering reasons. These paths can be strict or loose paths, 747 depending on the compactness required of the instruction list and the 748 degree of autonomy granted to the network, for example to support 749 Equal Cost Multipath load-balancing (ECMP) [RFC2992]. 751 With SR, a path needs to be dynamically created through a set of 752 segments by simply specifying the Segment Identifiers (SIDs), i.e. 753 instructions rooted at a particular point in the network. Thus if a 754 path is to be provisioned from some ingress point A to some egress 755 point B in the underlay, A is provided with a SID list from A to B 756 and instructions on how to identify the packets to which the SID list 757 is to be prepended. 759 By encoding the state in the packet, as is done in Segment Routing, 760 per-path state is transitioned out of the network. 762 However, there are a number of limitations in current SR, which limit 763 its applicability to enhanced VPNs: 765 o Segments are shared between different VPNs paths 767 o There is no reservation of bandwidth 769 o There is limited differentiation in the data plane. 771 Thus some extensions to SR are needed to provide isolation between 772 different enhanced VPNs. This can be achieved by including a finer 773 granularity of state in the network in anticipation of its future use 774 by authorized services. We therefore need to evaluate the balance 775 between this additional state and the performance delivered by the 776 network. 778 With current segment routing, the instructions are used to specify 779 the nodes and links to be traversed. However, in order to achieve 780 the required isolation between different services, new instructions 781 can be created which can be prepended to a packet to steer it through 782 specific network resources and functions. 784 Traditionally an SR traffic engineered path operates with a 785 granularity of a link with hints about priority provided through the 786 use of the traffic class (TC) field in the header. However to 787 achieve the latency and isolation characteristics that are sought by 788 the enhanced VPN users, steering packets through specific queues and 789 resources will likely be required. The extent to which these needs 790 can be satisfied through existing QoS mechanisms is to be determined. 791 What is clear is that a fine control of which services wait for 792 which, with a fine granularity of queue management policy is needed. 794 Note that the concept of a queue is a useful abstraction for many 795 types of underlay mechanism that may be used to provide enhanced 796 isolation and latency support. 798 From the perspective of the control plane, and from the perspective 799 of the segment routing, the method of steering a packet to a queue 800 that provides the required properties is an abstraction that hides 801 the details of the underlying implementation. How the queue 802 satisfies the requirement is implementation specific and is 803 transparent to the control plane and data plane mechanisms used. 804 Thus, for example, a FlexE channel, or a time sensitive networking 805 packet scheduling slot are abstracted to the same concept and bound 806 to the data plane in a common manner. 808 We can also introduce such fine grained packet steering by specifying 809 the queues through an SR instruction list. Thus new SR instructions 810 may be created to specify not only which resources are traversed, but 811 in some cases how they are traversed. For example, it may be 812 possible to specify not only the queue to be used but the policy to 813 be applied when enqueuing and dequeuing. 815 This concept could be further generalized, since as well as queuing 816 to the output port of a router, it is possible to consider queuing 817 data to any resource, for example: 819 o A network processor unit (NPU) 821 o A central processing unit (CPU) Core 823 o A Look-up engine 825 Both SR-MPLS and SRv6 are candidate network layer technologies for 826 enhanced VPN. In some cases they can be supported by DetNet to meet 827 the packet loss, delay and jitter requirement of particular service. 828 However, currently the "pure" IP variant of DetNet 829 [I-D.ietf-detnet-dp-sol-ip] does not support the Packet Replication, 830 Elimination, and Re-ordering (PREOF) [I-D.ietf-detnet-architecture] 831 functions. How to provide the DetNet enhanced delivery in an SRv6 832 environment needs further study. 834 4.3. Control Plane 836 Enhanced VPN would likely be based on a hybrid control mechanism, 837 which takes advantage of the logically centralized controller for on- 838 demand provisioning and global optimization, whilst still relies on 839 distributed control plane to provide scalability, high reliability, 840 fast reaction, automatic failure recovery etc. Extension and 841 optimization to the distributed control plane is needed to support 842 the enhanced properties of VPN+. 844 RSVP-TE provides the signaling mechanism of establishing a TE-LSP 845 with end-to-end resource reservation. It can be used to bind the VPN 846 to specific network resource allocated within the underlay, but there 847 are the above mentioned scalability concerns. 849 SR does not have the capability of signaling the resource reservation 850 along the path, nor do its currently specified distributed link state 851 routing protocols. On the other hand, the SR approach provides a way 852 of efficiently binding the network underlay and the enhanced VPN 853 overlay, as it reduces the amount of state to be maintained in the 854 network. An SR-based approach with per-slice resource reservation 855 can easily create dedicated SR network slices, and the VPN services 856 can be bound to a particular SR network slice. A centralized 857 controller can perform resource planning and reservation from the 858 controller's point of view, but this does not ensure resource 859 reservation is actually done in the network nodes. Thus, if a 860 distributed control plane is needed, either in place of an SDN 861 controller or as an assistant to it, the design of the control system 862 needs to ensure that resources are uniquely allocated in the network 863 nodes for the correct service, and not allocated to multiple services 864 causing unintended resource conflict. 866 4.4. Management Plane 868 The management plane mechanisms for enhanced VPN can be based on the 869 VPN service models as defined in [RFC8299] 870 [I-D.ietf-l2sm-l2vpn-service-model], possible augmentations and 871 extensions to these models may be needed, which is out of the scope 872 of this document. 874 Abstraction and Control of Traffic Engineered Networks (ACTN) 875 [I-D.ietf-teas-actn-framework] specifies the SDN based architecture 876 for the control of TE networks. The ACTN related data models such as 877 [I-D.ietf-teas-actn-vn-yang] and 878 [I-D.lee-teas-te-service-mapping-yang] can be applicable in the 879 provisioning of enhanced VPN service. The details are described in 880 [I-D.lee-rtgwg-actn-applicability-enhanced-vpn]. 882 5. Scalability Considerations 884 Enhanced VPN provides the performance guaranteed services in packet 885 networks, with the cost of introducing necessary additional states 886 into the network. There are at least three ways of adding the state 887 needed for VPN+: 889 o Introduce the complete state into the packet, as is done in SR. 890 This allows the controller to specify the detailed series of 891 forwarding and processing instructions for the packet as it 892 transits the network. The cost of this is an increase in the 893 packet header size. The cost is also that systems will have 894 capabilities enabled in case they are called upon by a service. 895 This is a type of latent state, and increases as we more precisely 896 specify the path and resources that need to be exclusively 897 available to a VPN. 899 o Introduce the state to the network. This is normally done by 900 creating a path using RSVP-TE, which can be extended to introduce 901 any element that needs to be specified along the path, for example 902 explicitly specifying queuing policy. It is of course possible to 903 use other methods to introduce path state, such as via a Software 904 Defined Network (SDN) controller, or possibly by modifying a 905 routing protocol. With this approach there is state per path per 906 path characteristic that needs to be maintained over its life- 907 cycle. This is more state than is needed using SR, but the packet 908 are shorter. 910 o Provide a hybrid approach based on using binding SIDs to create 911 path fragments, and bind them together with SR. 913 Dynamic creation of a VPN path using SR requires less state 914 maintenance in the network core at the expense of larger VPN headers 915 on the packet. The packet size can be lower if a form of loose 916 source routing is used (using a few nodal SIDs), and it will be lower 917 if no specific functions or resource on the routers are specified. 918 Reducing the state in the network is important to enhanced VPN, as it 919 requires the overlay to be more closely integrated with the underlay 920 than with traditional VPNs. This tighter coupling would normally 921 mean that more state needed to be created and maintained in the 922 network, as the state about fine granularity processing would need to 923 be loaded and maintained in the routers. However, a segment routed 924 approach allows much of this state to be spread amongst the network 925 ingress nodes, and transiently carried in the packets as SIDs. 927 These approaches are for further study. 929 5.1. Maximum Stack Depth of SR 931 One of the challenges with SR is the stack depth that nodes are able 932 to impose on packets [I-D.ietf-isis-segment-routing-msd]. This leads 933 to a difficult balance between adding state to the network and 934 minimizing stack depth, or minimizing state and increasing the stack 935 depth. 937 5.2. RSVP Scalability 939 The traditional method of creating a resource allocated path through 940 an MPLS network is to use the RSVP protocol. However there have been 941 concerns that this requires significant continuous state maintenance 942 in the network. There are ongoing works to improve the scalability 943 of RSVP-TE LSPs in the control plane [RFC8370]. 945 There is also concern at the scalability of the forwarder footprint 946 of RSVP as the number of paths through an LSR grows 947 [I-D.sitaraman-mpls-rsvp-shared-labels] proposes to address this by 948 employing SR within a tunnel established by RSVP-TE. 950 6. OAM Considerations 952 A study of OAM in SR networks has been documented in [RFC8403]. 954 The enhanced VPN OAM design needs to consider the following 955 requirements: 957 o Instrumentation of the underlay so that the network operator can 958 be sure that the resources committed to a tenant are operating 959 correctly and delivering the required performance. 961 o Instrumentation of the overlay by the tenant. This is likely to 962 be transparent to the network operator and to use existing 963 methods. Particular consideration needs to be given to the need 964 to verify the isolation and the various committed performance 965 characteristics. 967 o Instrumentation of the overlay by the network provider to 968 proactively demonstrate that the committed performance is being 969 delivered. This needs to be done in a non-intrusive manner, 970 particularly when the tenant is deploying a performance sensitive 971 application 973 o Verification of the conformity of the path to the service 974 requirement. This may need to be done as part of a commissioning 975 test. 977 These issues will be discussed in a future version of this document. 979 7. Enhanced Resiliency 981 Each enhanced VPN has a life-cycle, and needs modification during 982 deployment as the needs of its tenant change. Additionally, as the 983 network as a whole evolves, there will need to be garbage collection 984 performed to consolidate resources into usable quanta. 986 Systems in which the path is imposed such as SR, or some form of 987 explicit routing tend to do well in these applications, because it is 988 possible to perform an atomic transition from one path to another. 989 This is a single action by the head-end changes the path without the 990 need for coordinated action by the routers along the path. However, 991 implementations and the monitoring protocols need to make sure that 992 the new path is up and meet the required SLA before traffic is 993 transitioned to it. It is possible for deadlocks arise as a result 994 of the network becoming fragmented over time, such that it is 995 impossible to create a new path or modify a existing path without 996 impacting the SLA of other paths. Resolution of this situation is as 997 much a commercial issue as it is a technical issue and is outside the 998 scope of this document. 1000 There are however two manifestations of the latency problem that are 1001 for further study in any of these approaches: 1003 o The problem of packets overtaking one and other if a path latency 1004 reduces during a transition. 1006 o The problem of the latency transient in either direction as a path 1007 migrates. 1009 There is also the matter of what happens during failure in the 1010 underlay infrastructure. Fast reroute is one approach, but that 1011 still produces a transient loss with a normal goal of rectifying this 1012 within 50ms [RFC5654] . An alternative is some form of N+1 delivery 1013 such as has been used for many years to support protection from 1014 service disruption. This may be taken to a different level using the 1015 techniques proposed by the IETF deterministic network work with 1016 multiple in-network replication and the culling of later packets 1017 [I-D.ietf-detnet-architecture]. 1019 In addition to the approach used to protect high priority packets, 1020 consideration has to be given to the impact of best effort traffic on 1021 the high priority packets during a transient. Specifically if a 1022 conventional re-convergence process is used there will inevitably be 1023 micro-loops and whilst some form of explicit routing will protect the 1024 high priority traffic, lower priority traffic on best effort shortest 1025 paths will micro-loop without the use of a loop prevention 1026 technology. To provide the highest quality of service to high 1027 priority traffic, either this traffic must be shielded from the 1028 micro-loops, or micro-loops must be prevented. 1030 8. Security Considerations 1032 All types of virtual network require special consideration to be 1033 given to the isolation between the tenants. In this regard enhanced 1034 VPNs neither introduce, no experience a greater security risk than 1035 another VPN of the same base type. However, in an enhanced virtual 1036 network service the isolation requirement needs to be considered. If 1037 a service requires a specific latency then it can be damaged by 1038 simply delaying the packet through the activities of another tenant. 1039 In a network with virtual functions, depriving a function used by 1040 another tenant of compute resources can be just as damaging as 1041 delaying transmission of a packet in the network. The measures to 1042 address these dynamic security risks must be specified as part to the 1043 specific solution. 1045 9. IANA Considerations 1047 There are no requested IANA actions. 1049 10. Acknowledgements 1051 The authors would like to thank Charlie Perkins, James N Guichard and 1052 Adrian Farrel for their review and valuable comments. 1054 11. References 1056 11.1. Normative References 1058 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 1059 Requirement Levels", BCP 14, RFC 2119, 1060 DOI 10.17487/RFC2119, March 1997, 1061 . 1063 11.2. Informative References 1065 [BBF-SD406] 1066 "BBF SD-406: End-to-End Network Slicing", 2016, 1067 . 1070 [DETNET] "Deterministic Networking", March , 1071 . 1073 [FLEXE] "Flex Ethernet Implementation Agreement", March 2016, 1074 . 1077 [I-D.ietf-detnet-architecture] 1078 Finn, N., Thubert, P., Varga, B., and J. Farkas, 1079 "Deterministic Networking Architecture", draft-ietf- 1080 detnet-architecture-08 (work in progress), September 2018. 1082 [I-D.ietf-detnet-dp-sol-ip] 1083 Korhonen, J. and B. Varga, "DetNet IP Data Plane 1084 Encapsulation", draft-ietf-detnet-dp-sol-ip-00 (work in 1085 progress), July 2018. 1087 [I-D.ietf-detnet-use-cases] 1088 Grossman, E., "Deterministic Networking Use Cases", draft- 1089 ietf-detnet-use-cases-19 (work in progress), October 2018. 1091 [I-D.ietf-isis-segment-routing-msd] 1092 Tantsura, J., Chunduri, U., Aldrin, S., and L. Ginsberg, 1093 "Signaling MSD (Maximum SID Depth) using IS-IS", draft- 1094 ietf-isis-segment-routing-msd-19 (work in progress), 1095 October 2018. 1097 [I-D.ietf-l2sm-l2vpn-service-model] 1098 Wen, B., Fioccola, G., Xie, C., and L. Jalil, "A YANG Data 1099 Model for L2VPN Service Delivery", draft-ietf-l2sm-l2vpn- 1100 service-model-10 (work in progress), April 2018. 1102 [I-D.ietf-teas-actn-framework] 1103 Ceccarelli, D. and Y. Lee, "Framework for Abstraction and 1104 Control of Traffic Engineered Networks", draft-ietf-teas- 1105 actn-framework-15 (work in progress), May 2018. 1107 [I-D.ietf-teas-actn-vn-yang] 1108 Lee, Y., Dhody, D., Ceccarelli, D., Bryskin, I., Yoon, B., 1109 Wu, Q., and P. Park, "A Yang Data Model for ACTN VN 1110 Operation", draft-ietf-teas-actn-vn-yang-02 (work in 1111 progress), September 2018. 1113 [I-D.lee-rtgwg-actn-applicability-enhanced-vpn] 1114 King, D., Lee, Y., Tantsura, J., Wu, Q., and D. 1115 Ceccarelli, "Applicability of Abstraction and Control of 1116 Traffic Engineered Networks (ACTN) to Enhanced VPN", 1117 draft-lee-rtgwg-actn-applicability-enhanced-vpn-03 (work 1118 in progress), July 2018. 1120 [I-D.lee-teas-te-service-mapping-yang] 1121 Lee, Y., Dhody, D., Ceccarelli, D., Tantsura, J., 1122 Fioccola, G., and Q. Wu, "Traffic Engineering and Service 1123 Mapping Yang Model", draft-lee-teas-te-service-mapping- 1124 yang-12 (work in progress), October 2018. 1126 [I-D.sitaraman-mpls-rsvp-shared-labels] 1127 Sitaraman, H., Beeram, V., Parikh, T., and T. Saad, 1128 "Signaling RSVP-TE tunnels on a shared MPLS forwarding 1129 plane", draft-sitaraman-mpls-rsvp-shared-labels-03 (work 1130 in progress), December 2017. 1132 [NGMN-NS-Concept] 1133 "NGMN NS Concept", 2016, . 1137 [RFC2475] Blake, S., Black, D., Carlson, M., Davies, E., Wang, Z., 1138 and W. Weiss, "An Architecture for Differentiated 1139 Services", RFC 2475, DOI 10.17487/RFC2475, December 1998, 1140 . 1142 [RFC2764] Gleeson, B., Lin, A., Heinanen, J., Armitage, G., and A. 1143 Malis, "A Framework for IP Based Virtual Private 1144 Networks", RFC 2764, DOI 10.17487/RFC2764, February 2000, 1145 . 1147 [RFC2992] Hopps, C., "Analysis of an Equal-Cost Multi-Path 1148 Algorithm", RFC 2992, DOI 10.17487/RFC2992, November 2000, 1149 . 1151 [RFC3209] Awduche, D., Berger, L., Gan, D., Li, T., Srinivasan, V., 1152 and G. Swallow, "RSVP-TE: Extensions to RSVP for LSP 1153 Tunnels", RFC 3209, DOI 10.17487/RFC3209, December 2001, 1154 . 1156 [RFC3758] Stewart, R., Ramalho, M., Xie, Q., Tuexen, M., and P. 1157 Conrad, "Stream Control Transmission Protocol (SCTP) 1158 Partial Reliability Extension", RFC 3758, 1159 DOI 10.17487/RFC3758, May 2004, 1160 . 1162 [RFC3931] Lau, J., Ed., Townsley, M., Ed., and I. Goyret, Ed., 1163 "Layer Two Tunneling Protocol - Version 3 (L2TPv3)", 1164 RFC 3931, DOI 10.17487/RFC3931, March 2005, 1165 . 1167 [RFC3985] Bryant, S., Ed. and P. Pate, Ed., "Pseudo Wire Emulation 1168 Edge-to-Edge (PWE3) Architecture", RFC 3985, 1169 DOI 10.17487/RFC3985, March 2005, 1170 . 1172 [RFC4364] Rosen, E. and Y. Rekhter, "BGP/MPLS IP Virtual Private 1173 Networks (VPNs)", RFC 4364, DOI 10.17487/RFC4364, February 1174 2006, . 1176 [RFC4448] Martini, L., Ed., Rosen, E., El-Aawar, N., and G. Heron, 1177 "Encapsulation Methods for Transport of Ethernet over MPLS 1178 Networks", RFC 4448, DOI 10.17487/RFC4448, April 2006, 1179 . 1181 [RFC4594] Babiarz, J., Chan, K., and F. Baker, "Configuration 1182 Guidelines for DiffServ Service Classes", RFC 4594, 1183 DOI 10.17487/RFC4594, August 2006, 1184 . 1186 [RFC4664] Andersson, L., Ed. and E. Rosen, Ed., "Framework for Layer 1187 2 Virtual Private Networks (L2VPNs)", RFC 4664, 1188 DOI 10.17487/RFC4664, September 2006, 1189 . 1191 [RFC4719] Aggarwal, R., Ed., Townsley, M., Ed., and M. Dos Santos, 1192 Ed., "Transport of Ethernet Frames over Layer 2 Tunneling 1193 Protocol Version 3 (L2TPv3)", RFC 4719, 1194 DOI 10.17487/RFC4719, November 2006, 1195 . 1197 [RFC5654] Niven-Jenkins, B., Ed., Brungard, D., Ed., Betts, M., Ed., 1198 Sprecher, N., and S. Ueno, "Requirements of an MPLS 1199 Transport Profile", RFC 5654, DOI 10.17487/RFC5654, 1200 September 2009, . 1202 [RFC7209] Sajassi, A., Aggarwal, R., Uttaro, J., Bitar, N., 1203 Henderickx, W., and A. Isaac, "Requirements for Ethernet 1204 VPN (EVPN)", RFC 7209, DOI 10.17487/RFC7209, May 2014, 1205 . 1207 [RFC8299] Wu, Q., Ed., Litkowski, S., Tomotaki, L., and K. Ogaki, 1208 "YANG Data Model for L3VPN Service Delivery", RFC 8299, 1209 DOI 10.17487/RFC8299, January 2018, 1210 . 1212 [RFC8370] Beeram, V., Ed., Minei, I., Shakir, R., Pacella, D., and 1213 T. Saad, "Techniques to Improve the Scalability of RSVP-TE 1214 Deployments", RFC 8370, DOI 10.17487/RFC8370, May 2018, 1215 . 1217 [RFC8402] Filsfils, C., Ed., Previdi, S., Ed., Ginsberg, L., 1218 Decraene, B., Litkowski, S., and R. Shakir, "Segment 1219 Routing Architecture", RFC 8402, DOI 10.17487/RFC8402, 1220 July 2018, . 1222 [RFC8403] Geib, R., Ed., Filsfils, C., Pignataro, C., Ed., and N. 1223 Kumar, "A Scalable and Topology-Aware MPLS Data-Plane 1224 Monitoring System", RFC 8403, DOI 10.17487/RFC8403, July 1225 2018, . 1227 [SFC] "Deterministic Networking", March , 1228 . 1230 [TS23501] "3GPP TS23.501", 2016, 1231 . 1234 [TS28530] "3GPP TS28.530", 2016, 1235 . 1238 [TSN] "Time-Sensitive Networking", March , 1239 . 1241 Authors' Addresses 1243 Jie Dong 1244 Huawei 1246 Email: jie.dong@huawei.com 1248 Stewart Bryant 1249 Huawei 1251 Email: stewart.bryant@gmail.com 1253 Zhenqiang Li 1254 China Mobile 1256 Email: lizhenqiang@chinamobile.com 1258 Takuya Miyasaka 1259 KDDI Corporation 1261 Email: ta-miyasaka@kddi.com