Internet Draft Loa Andersson, Ed. (Acreo AB) Category: Informational Lou Berger, Ed. (LabN) Expiration Date: January 13, 2010 Luyuan Fang, Ed. (Cisco) Nabil Bitar, Ed. (Verizon) July 13, 2009 MPLS-TP Control Plane Framework draft-abfb-mpls-tp-control-plane-framework-01.txt Status of this Memo This Internet-Draft is submitted to IETF in full conformance with the provisions of BCP 78 and BCP 79. By submitting this Internet-Draft, each author represents that any applicable patent or other IPR claims of which he or she is aware have been or will be disclosed, and any of which he or she becomes aware will be disclosed, in accordance with BCP 78 and BCP 79. Internet-Drafts are working documents of the Internet Engineering Task Force (IETF), its areas, and its working groups. Note that other groups may also distribute working documents as Internet- Drafts. Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress." The list of current Internet-Drafts can be accessed at http://www.ietf.org/1id-abstracts.html The list of Internet-Draft Shadow Directories can be accessed at http://www.ietf.org/shadow.html This Internet-Draft will expire on January 13, 2010. Copyright and License Notice Copyright (c) 2009 IETF Trust and the persons identified as the document authors. All rights reserved. This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents in effect on the date of publication of this document (http://trustee.ietf.org/license-info). Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Andersson, et al Informational [Page 1] Internet-Draft draft-abfb-mpls-tp-control-plane-framework-01.txt July 13, 2009 Abstract The MPLS Transport Profile (MPLS-TP) supports both static provisioning of transport paths via an NMS/OSS, and dynamic provisioning of transport paths via a control plane. This document provides the framework for MPLS-TP dynamic provisioning, and covers control plane signaling, routing, addressing, traffic engineering, path computation, and recovery in the event of network failures. The document focuses on the control of Label Switched Paths (LSPs) as the Pseudowire (PW) control plane is not modified by MPLS-TP. MPLS-TP uses GMPLS as the control plane for MPLS-TP LSPs. Backwards compatibility to MPLS is required. Management plane functions such as manual configuration, the initiation of LSP setup are out of scope of this document. Table of Contents 1 Introduction ........................................... 3 1.1 Conventions Used In This Document ...................... 3 1.2 Scope .................................................. 3 1.3 Basic Approach ......................................... 4 1.4 Reference Model ........................................ 5 2 Control plane requirements ............................. 8 2.1 Primary Requirements ................................... 8 2.2 MPLS-TP Framework Derived Requirements ................. 16 2.3 OAM Framework Derived Requirements ..................... 17 2.4 Security Requirements .................................. 19 3 Relationship of PWs and TE LSPs ........................ 19 4 TE LSPs ................................................ 19 4.1 General reuse of existing GMPLS control plane mechanisms ...19 4.1.1 In-Band and Out-Of-Band ................................ 20 4.1.2 Addressing ............................................. 21 4.2 Signaling .............................................. 21 4.3 Routing ................................................ 22 4.3.1 ISIS-TE/OSPF-TE routing in support of MPLS-TP .......... 23 4.3.2 TE link bundling ....................................... 25 4.4 OAM, MEP (hierarchy) configuration & control ........... 25 4.5 Traffic engineering and constraint-based path computation ..25 4.5.1 Relation to PCE ........................................ 26 4.6 Applicability .......................................... 26 4.7 Recovery ............................................... 26 4.7.1 E2E, segment ........................................... 26 4.7.2 P2P, P2MP .............................................. 27 4.7.3 Recovery Triggers ...................................... 27 4.8 Diffserv object usage in GMPLS (E-LSPs, L-LSPs) ........ 27 4.9 Management plane support ............................... 27 4.9.1 Management Plane / Control Plane Ownership Transfer .... 27 4.10 CP reference points (E-NNI, I-NNI, UNI) ................ 28 4.11 MPLS to MPLS-TP interworking ........................... 28 5 Pseudo Wires ........................................... 28 5.1 General reuse of existing PW control plane mechanisms .. 31 Andersson, et al Informational [Page 2] Internet-Draft draft-abfb-mpls-tp-control-plane-framework-01.txt July 13, 2009 5.2 Signaling .............................................. 31 5.3 Recovery (Redundancy) .................................. 31 6 Security Considerations ................................ 31 7 IANA Considerations .................................... 31 8 Acknowledgments ........................................ 31 9 References ............................................. 32 9.1 Normative References ................................... 32 9.2 Informative References ................................. 34 10 Authors' Addresses ..................................... 36 1. Introduction The MPLS Transport Profile (MPLS-TP) is being defined in a joint effort between the International Telecommunications Union (ITU) and the IETF. The requirements for MPLS-TP are defined in the requirements document, see [TP-REQ]. These requirements state that "A solution MUST be provided to support dynamic provisioning of MPLS- TP transport paths via a control plane." This document provides the framework for such dynamic provisioning. 1.1. Conventions Used In This Document The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in [RFC2119]. 1.2. Scope This document covers control plane related topics for MPLS-TP Label Switched Paths (LSPs) and Pseudowire (PW). The control plane requirements for MPLS-TP are defined in [TP-REQ]. These requirements defined the role of the control plane in MPLS-TP. In particular, Sections 2.4 and portions of the remainder of Section 2 of [TP-REQ] provide specific control plane requirements. The LSPs provided by MPLS-TP are used as a server layer for IP, MPLS and PWs, as well as other tunneled MPLS-TP LSPs. The PWs are used to carry client signal other than IP and MPLS. The relationship between pseudo wires and MPLS-TP LSPs is exactly the same as between pseudo wires and MPLS LSPs in a Packet switched network (PSN). The PW encapsulation over MPLS-TP LSPs used in MPLS-TP networks is the same Andersson, et al Informational [Page 3] Internet-Draft draft-abfb-mpls-tp-control-plane-framework-01.txt July 13, 2009 as for PWs over MPLS in an MPLS network. MPLS-TP also defines protection and restoration (or, collectively, recovery) functions. The MPLS-TP control plane provides methods to establish, remove and control MPLS-TP LSP and PW functions. This includes control of data plane, OAM and recovery functions. A general framework for MPLS-TP has been defined in [TP-FWK], and a survivability framework for MPLS-TP has been defined in [TP-SURVIVE]. These document scope the approaches and protocols that will be used as the foundation for MPLS-TP. Notably, Section 3.5 of [TP-FWK] scopes the IETF protocols that serve as the foundation of the MPLS-TP control plane. The PW control plane is based on the existing PW control plane, see [RFC4447], and the PW end-to-end (PWE3) architecture, see [RFC3985]. The LSP control plane is based on Generalized MPLS (GMPLS), see [RFC3945], which is built on MPLS-TE and its numerous extensions. [TP-SURVIVE] focuses on LSPs, and the protection functions that must be supported within MPLS-TP. It does not specify which control plane mechanisms are to be used. This document discusses the impact of MPLS-TP requirements on the signaling that is used to provision pseudo wires as specified in RFC4447. This document also discusses the impact of the MPLS-TP requirements on the GMPLS signaling and routing protocols that is used to provision MPLS-TP LSPs. 1.3. Basic Approach The basic approach taken in defining the MPLS-TP Control Plane framework is: 1) MPLS technology as defined by the IETF is the foundation for the MPLS Transport Profile. [Editor's note: should also be in primary TP documents.] 2) The data plane for MPLS and MPLS-TP is identical, i.e. any extensions defined for MPLS-TP is also applicable to MPLS. And the same encapsulation used for MPLS over any layer 2 network is also used for MPLS-TP. [Editor's note: should also be in primary TP documents.] 3) MPLS PWs are used as-is by MPLS-TP including the use of targeted-LDP for PW signaling [RFC4447], OSPF-TE, ISIS-TE or MP-BGP as they apply for Multi-Segment(MS)-PW routing. However, the PW can be encapsulated over an MPLS-TP LSP in (established using methods and procedures for MPLS-TP LSP establishment) in addition to the presently defined methods of carrying PWs over packet switched networks (PSNs). That is, the MPLS-TP domain is a packet switched network from PWE3 architecture aspect [RFC3985]. Andersson, et al Informational [Page 4] Internet-Draft draft-abfb-mpls-tp-control-plane-framework-01.txt July 13, 2009 4) The MPLS-TP LSP control plane builds on the GMPLS control plane as defined by the IETF for transport LSPs, the protocols within scope are RSVP-TE [RFC3473], OSPF-TE [RFC4203][RFC5392], and ISIS-TE [RFC5307][RFC5316]. ASON/ASTN signaling and routing requirements in the context of GMPLS can be found in [RFC4139] and [RFC4258]. 5) Existing IETF MPLS and GMPLS RFCs and evolving Working Group Internet-Drafts should be reused wherever possible. 6) If needed, extensions for the MPLS-TP control plane should first based on the existing and evolving IETF work, secondly based on work by other Standard bodies only when IETF decides that the work is out of IETF scope. New extensions may be defined otherwise. 7) Extensions to the GMPLS control plane may be required in order to fully automate MPLS-TP functions. 8) Control-plane software upgrades to existing equipment running MPLS is acceptable and expected. 9) It is permissible for functions present in the GMPLS control plane not to be used in MPLS-TP networks, e.g. the possibility to merge LSPs. 10) One possible use of the control plane is to configure, enable and empower OAM functionality; this will require extensions to existing control plane specifications. 11) MPLS-TP requirements are primarily defined in Section 2.4 and relevant portions of the remainder Section 2 of [TP-REQ]. 1.4. Reference Model The control plane reference model is based on the general MPLS-TP reference model as defined in MPLS-TP framework [TP-FWK]. Per MPLS-TP framework [TP-FWK], MPLS-TP control plane is based on GMPLS with RSVP-TE for LSP signaling and LDP for PW signaling. In both cases, OSPF-TE or ISIS-TE with GMPLS extensions is used for dynamic routing. From a service perspective, client interfaces are provided for both the PWs and LSPs. PW client interfaces are defined on an interface technology basis, e.g., Ethernet over PW [RFC4448]. In the context of MPLS-TP LSP, the client interface is expected to be provided via a UNI, [RFC4208]. As discussed in [TP-FWK], MPLS-TP also presumes an LSP NNI reference point. The MPLS-TP end-to-end control plane reference model is shown in Figure 1. It shows the control plane protocols used by MPLS-TP, as well as the UNI and NNI reference points. [Editor's note: need to add more explanatory text.] Andersson, et al Informational [Page 5] Internet-Draft draft-abfb-mpls-tp-control-plane-framework-01.txt July 13, 2009 |< ---- client signal (IP / MPLS / L2 / PW) ------------ >| |< --------- SP1 ----------- >|< ------- SP2 ------- >| |< ---------- MPLS-TP End to End PW ------------ >| |< -------- MPLS-TP End to End LSP --------- >| +---+ +---+ +---+ +---+ +---+ +---+ +---+ +---+ +---+ |CE1|-|-|PE1|--|P1 |--|P2 |--|PE2|-|-|PEa|--|Pa |--|PEb|-|-|CE2| +---+ +---+ +---+ +---+ +---+ +---+ +---+ +---+ +---+ UNI NNI UNI TE-RTG |< ---------------- >|< --- >|< ---------- >| RSVP-TE LDP |< --------------------------------------- >| Figure 1. End-to-End MPLS-TP Control Plane Reference Model Legend: CE: Customer Edge Client signal: defined in MPLS-TP Requirements L2: Any layer 2 signal that may be carried over a PW, e.g. Ethernet. NNI: Network to Network Interface PE: Provider Edge SP: Service Provider TE-RTG: OSPF-TE or ISIS-TE UNI: User to Network Interface Figure 2 adds three hierarchical LSP segments, labeled as "H- LSPs". These segments are present to support OAM and MEPs within each provider and across the inter-provider NNI. The MEPs are used to collect performance information and support OAM triggered survivability schemes as discussed in [TP-SURVIVE], and each H-LSP may be protected using any of the schemes discussed in [TP-SURVIVE]. Andersson, et al Informational [Page 6] Internet-Draft draft-abfb-mpls-tp-control-plane-framework-01.txt July 13, 2009 |< ------- client signal (IP / MPLS / L2 / PW) ------ >| |< -------- SP1 ----------- >|< ------- SP2 ----- >| |< ----------- MPLS-TP End to End PW -------- >| |< ------- MPLS-TP End to End LSP ------- >| |< -- H-LSP1 ---- >|<-H-LSP2->|<- H-LSP3 ->| +---+ +---+ +---+ +---+ +---+ +---+ +---+ +---+ +---+ |CE1|-|-|PE1|--|P1 |--|P2 |--|PE2|-|-|PEa|--|Pa |--|PEb|-|-|CE2| +---+ +---+ +---+ +---+ +---+ +---+ +---+ +---+ +---+ UNI NNI UNI ..... ..... ..... ......... ......... ..... ..... |MEP|-|MIP|-|MIP|-|MEP|MEP|-|MEP|MEP|-|MIP|-|MEP| ''''' ''''' ''''' ''''''''' ''''''''' ''''' ''''' TE-RTG |< -- >|< -- >|< -- >||< -- >||< -- >|< -- >| RSVP-TE (within the MPLS-TP domain) TE-RTG |< ---------------- >|< ---- >|< --------- >| RSVP-TE LDP |< --------------------------------------- >| Figure 2. MPLS-TP Control Plane Reference Model with OAM Legend: CE: Customer Edge Client signal: defined in MPLS-TP Requirements L2: Any layer 2 signal that may be carried over a PW, e.g. Ethernet. H-LSP: Hierarchical LSP MEP: Maintenance end points MIP: Maintenance intermediate points NNI: Network to Network Interface PE: Provider Edge SP: Service Provider TE-RTG: OSPF-TE or ISIS-TE While not shown in the Figures above, it is worth noting that the MPLS-TP control plane must support the addressing separation and independence between the data, control and management planes as shown in Figure 3 of [TP-FWK]. Address separation between the planes is already included in GMPLS. Andersson, et al Informational [Page 7] Internet-Draft draft-abfb-mpls-tp-control-plane-framework-01.txt July 13, 2009 2. Control plane requirements The requirements for the MPLS-TP control plane are derived from the MPLS-TP requirements and framework documents, specifically [TP-REQ], [TP-FWK], [TP-OAM-REQ], [TP-OAM], and [TP-SURVIVE]. The requirements are summarized in this section, but do not replace those documents. If there are differences between this section and those documents, those documents shall be considered authoritative. 2.1. Primary Requirements These requirements are based on Section 2 [TP-REQ]: 1. Any new functionality that is defined to fulfill the requirements for MPLS-TP must be agreed within the IETF through the IETF consensus process as per [RFC4929]. 2. The MPLS-TP control plane design should as far as reasonably possible re-use existing MPLS standards. 3. The MPLS-TP control plane must be able to interoperate with existing IETF MPLS and PWE3 control planes where appropriate. 4. The MPLS-TP control plane must be sufficiently well-defined that interworking equipment supplied by multiple vendors will be possible both within a single domain, and between domains. 5. The MPLS-TP control plane must support a connection- oriented packet switching model with traffic engineering capabilities that allow deterministic control of the use of network resources. 6. The MPLS-TP control plane must support traffic engineered point-to-point (P2P) and point-to-multipoint (P2MP) transport paths. 7. The MPLS-TP control plane must support unidirectional, associated bidirectional and co-routed bidirectional point-to- point transport paths. 8. The MPLS-TP control plane must support unidirectional point-to- multipoint transport paths. 9. All nodes (i.e., ingress, egress and intermediate) must be aware about the pairing relationship of the forward and the backward directions belonging to the same co-routed bidirectional transport path. Andersson, et al Informational [Page 8] Internet-Draft draft-abfb-mpls-tp-control-plane-framework-01.txt July 13, 2009 10. Edge nodes (i.e., ingress and egress) must be aware about the pairing relationship of the forward and the backward directions belonging to the same associated bidirectional transport path. 11. Transit nodes should be aware about the pairing relationship of the forward and the backward directions belonging to the same associated bidirectional transport path. 12. The MPLS-TP control plane must support bidirectional transport paths with symmetric bandwidth requirements, i.e. the amount of reserved bandwidth is the same in the forward and backward directions. 13. The MPLS-TP control plane must support bidirectional transport paths with asymmetric bandwidth requirements, i.e. the amount of reserved bandwidth differs in the forward and backward directions. 14. The MPLS-TP control plane must support the logical separation of the control and management planes from the data plane. 15. The MPLS-TP control plane must support the physical separation of the control and management planes from the data plane. 16. A control plane must be defined to support dynamic provisioning and restoration of MPLS-TP transport paths, but its use is a network operator's choice. 17. A control plane must not be required to support the static provisioning of MPLS-TP transport paths. 18. The MPLS-TP control plane must permit the co-existence of statically and dynamically provisioned/managed MPLS-TP transport paths within the same layer network or domain. 19. The MPLS-TP control plane should be operable in a way that is similar to the way the control plane operates in other transport layer technologies. 20. The MPLS-TP control plane must avoid or minimize traffic impact (e.g. packet delay, reordering and loss) during network reconfiguration. 21. The MPLS-TP control plane must work across multiple homogeneous domains. 22. The MPLS-TP control plane should work across multiple non- homogeneous domains. Andersson, et al Informational [Page 9] Internet-Draft draft-abfb-mpls-tp-control-plane-framework-01.txt July 13, 2009 23. The MPLS-TP control plane must not dictate any particular physical or logical topology. 24. The MPLS-TP control plane must include support of ring topologies which may be deployed with arbitrarily interconnection, support rings of at least 16 nodes. 25. The MPLS-TP control plane must support scale gracefully to support a large number of transport paths, nodes and links. That is it must be able to scale at least as well as control planes in existing transport technologies with growing and increasingly complex network topologies as well as with increasing bandwidth demands, number of customers, and number of services. 26. The MPLS-TP control plane should not provision transport paths which contain forwarding loops. 27. The MPLS-TP control plane must support multiple client layers. (e.g. MPLS-TP, IP, MPLS, Ethernet, ATM, FR, etc.) 28. The MPLS-TP control plane must provide a generic and extensible solution to support the transport of MPLS-TP transport paths over one or more server layer networks (such as MPLS-TP, Ethernet, SONET/SDH, OTN, etc.). Requirements for bandwidth management within a server layer network are outside the scope of this document. 29. In an environment where an MPLS-TP layer network is supporting a client layer network, and the MPLS-TP layer network is supported by a server layer network then the control plane operation of the MPLS-TP layer network must be possible without any dependencies on the server or client layer network. 30. The MPLS-TP control plane must allow for the transport of a client MPLS or MPLS-TP layer network over a server MPLS or MPLS-TP layer network. 31. The MPLS-TP control plane must allow the operation the layers of a multi-layer network that includes an MPLS-TP layer autonomously. 32. The MPLS-TP control plane must allow the hiding of MPLS-TP layer network addressing and other information (e.g. topology) from client layer networks. However, it should be possible, at the option of the operator, to leak a limited amount of summarized information (such as SRLGs or reachability) between layers. Andersson, et al Informational [Page 10] Internet-Draft draft-abfb-mpls-tp-control-plane-framework-01.txt July 13, 2009 33. The MPLS-TP control plane must allow for the identification of a transport path on each link within and at the destination (egress) of the transport network. 34. The MPLS-TP control plane must allow for P2MP capable server (sub-)layers. 35. The MPLS-TP control plane must be extensible in order to accommodate new types of client layer networks and services. 36. The MPLS-TP control plane should support the reserved bandwidth associated with a transport path to be increased without impacting the existing traffic on that transport path provided enough resources are available. 37. The MPLS-TP control plane should support the reserved bandwidth of a transport path to be decreased without impacting the existing traffic on that transport path, provided that the level of existing traffic is smaller than the reserved bandwidth following the decrease. 38. The MPLS-TP control plane must support an unambiguous and reliable means of distinguishing users' (client) packets from MPLS-TP control packets (e.g. control plane, management plane, OAM and protection switching packets). 39. The control plane for MPLS-TP must fit within the ASON architecture. The ITU-T has defined an architecture for Automatically Switched Optical Networks (ASON) in G.8080 [ITU.G8080.2006] and G.8080 Amendment 1 [ITU.G8080.2008]. An interpretation of the ASON signaling and routing requirements in the context of GMPLS can be found in [RFC4139] and [RFC4258]. 40. The MPLS-TP control plane must support control plane topology and data plane topology independence. 41. A failure of the MPLS-TP control plane must not interfere with the deliver of service or recovery of established transport paths. 42. The MPLS-TP control plane must be able to operate independent of any particular client or server layer control plane. 43. The MPLS-TP control plane should support, but not require, an integrated control plane encompassing MPLS-TP together with its server and client layer networks when these layer networks belong to the same administrative domain. Andersson, et al Informational [Page 11] Internet-Draft draft-abfb-mpls-tp-control-plane-framework-01.txt July 13, 2009 44. The MPLS-TP control plane must support configuration of protection functions and any associated maintenance (OAM) functions. 45. The MPLS-TP control plane must support the configuration and modification of OAM maintenance points as well as the activation/deactivation of OAM when the transport path or transport service is established or modified. 46. The MPLS-TP control plane must capable of restarting and relearning its previous state without impacting forwarding. 47. The MPLS-TP control plane must provide a mechanism for dynamic ownership transfer of the control of MPLS-TP transport paths from the management plane to the control plane and vice versa. The number of reconfigurations required in the data plane must be minimized (preferably no data plane reconfiguration will be required). 48. The MPLS-TP control plane must support protection and restoration mechanisms, i.e., recovery. Note that the MPLS-TP Survivability Framework document, [TP- SURVIVE], provides additional useful information related to recovery. 49. The MPLS-TP control plane mechanisms should be identical (or as similar as possible) to those already used in existing transport networks to simplify implementation and operations. However, this must not override any other requirement. 50. The MPLS-TP control plane mechanisms used for P2P and P2MP recovery should be identical to simplify implementation and operation. However, this must not override any other requirement. 51. The MPLS-TP control plane must support recovery mechanisms that are applicable at various levels throughout the network including support for link, transport path, segment, concatenated segment and end to end recovery. 52. The MPLS-TP control plane must support recovery paths that meet the SLA protection objectives of the service. Including: a. Guarantee 50ms recovery times from the moment of fault detection in networks with spans less than 1200 km. b. Protection of up to 100% of the traffic on the protected path. Andersson, et al Informational [Page 12] Internet-Draft draft-abfb-mpls-tp-control-plane-framework-01.txt July 13, 2009 c. Recovery must meet SLA requirements over multiple domains. 53. The MPLS-TP control plane should support per transport path Recovery objective. 54. The MPLS-TP control plane must support recovery mechanisms that are applicable to any topology. 55. The MPLS-TP control plane must operate in synergy with (including coordination of timing/timer settings) the recovery mechanisms present in any client or server transport networks (for example, Ethernet, SDH, OTN, WDM) to avoid race conditions between the layers. 56. The MPLS-TP control plane must support recovery and reversion mechanisms that prevent frequent operation of recovery in the event of an intermittent defect. 57. The MPLS-TP control plane must support revertive and non- revertive protection behavior. 58. The MPLS-TP control plane must support 1+1 bidirectional protection for P2P transport paths. 59. The MPLS-TP control plane must support 1+1 unidirectional protection for P2P transport paths. 60. The MPLS-TP control plane must support 1+1 unidirectional protection for P2MP transport paths. 61. The MPLS-TP control plane must support the ability to share protection resources amongst a number of transport paths. 62. The MPLS-TP control plane must support 1:n bidirectional protection for P2P transport paths, and this should be the default for 1:n protection. 63. The MPLS-TP control plane must support 1:n unidirectional protection for P2MP transport paths. 64. The MPLS-TP control plane may support 1:n unidirectional protection for P2P transport paths. 65. The MPLS-TP control plane may support extra-traffic. 66. The MPLS-TP control plane should support 1:n (including 1:1) shared mesh recovery. Andersson, et al Informational [Page 13] Internet-Draft draft-abfb-mpls-tp-control-plane-framework-01.txt July 13, 2009 67. The MPLS-TP control plane must support sharing of protection resources such that protection paths that are known not to be required concurrently can share the same resources. 68. The MPLS-TP control plane must support the sharing of resources between a restoration transport path and the transport path being replaced. 69. The MPLS-TP control plane must support restoration priority so that an implementation can determine the order in which transport paths should be restored. 70. The MPLS-TP control plane must support preemption priority in order to allow restoration to displace other transport paths in the event of resource constraints. 71. The MPLS-TP control plane may support revertive and non- revertive restoration behavior. 72. The MPLS-TP control plane must support recovery being triggered by physical (lower) layer fault indications. 73. The MPLS-TP control plane must support recovery being triggered by OAM. 74. The MPLS-TP control plane must support management plane recovery triggers (e.g., forced switch, etc.). 75. The MPLS-TP control plane must support the differentiation of administrative recovery actions from recovery actions initiated by other triggers. 76. The MPLS-TP control plane should support control plane restoration triggers (e.g., forced switch, etc.). 77. The MPLS-TP control plane must support priority logic to negotiate and accommodate coexisting requests (i.e., multiple requests) for protection switching (e.g., administrative requests and requests due to link/node failures). 78. The MPLS-TP control plane must support the relationships of protection paths and protection-to-working paths (sometimes known as protection groups). 79. The MPLS-TP control plane must support pre-calculation of recovery paths. 80. The MPLS-TP control plane must support pre-provisioning of recovery paths. Andersson, et al Informational [Page 14] Internet-Draft draft-abfb-mpls-tp-control-plane-framework-01.txt July 13, 2009 81. The MPLS-TP control plane must support the external commands defined in [RFC4427]. External controls overruled by higher priority requests (e.g., administrative requests and requests due to link/node failures) or unable to be signaled to the remote end (e.g. because of a protection state coordination fail) must be dropped. 82. The MPLS-TP control plane must permit the testing and validation of the integrity of the protection/recovery transport path. 83. The MPLS-TP control plane must permit the testing and validation of protection/ restoration mechanisms without triggering the actual protection/restoration. 84. The MPLS-TP control plane must permit the testing and validation of protection/ restoration mechanisms while the working path is in service. 85. The MPLS-TP control plane must permit the testing and validation of protection/ restoration mechanisms while the working path is out of service. 86. The MPLS-TP control plane must support the establishment and maintenance of all recovery entities and functions. 87. The MPLS-TP control plane must support signaling of recovery administrative control. 88. The MPLS-TP control plane must support protection state coordination (PSC). Since control plane network topology is independent from the data plane network topology, the PSC supported by the MPLS-TP control plane may run on resources different than the data plane resources handled within the recovery mechanism (e.g. backup). 89. The MPLS-TP control plane may support recovery mechanisms that are optimized for specific network topologies. These mechanisms must be interoperable with the mechanisms defined for arbitrary topology (mesh) networks to enable protection of end-to-end transport paths. 90. (Editor's note: inclusion of ring specific control plane requirements are TBD, See Section 2.8.6.1. of [TP-REQ]) 91. The MPLS-TP control plane must include support for differentiated services and different traffic types with traffic class separation associated with different traffic. Andersson, et al Informational [Page 15] Internet-Draft draft-abfb-mpls-tp-control-plane-framework-01.txt July 13, 2009 92. The MPLS-TP control plane must support the provisioning of services that provide a guaranteed of Service Level Specifications (SLS), with support for hard and relative end- to-end bandwidth guarantees. [Editor's note: add reference to definition of hard and relative guarantees] 93. The MPLS-TP control plane must support the provisioning of services which are sensitive to jitter and delay. 2.2. MPLS-TP Framework Derived Requirements The following additional requirements are based on [TP-FWK]: 94. Per-packet equal cost multi-path (ECMP) load balancing is not applicable to MPLS-TP. 95. Penultimate hop popping (PHP) is disabled on MPLS-TP LSPs by default. The applicability of PHP to both MPLS-TP LSPs and MPLS networks general providing packet transport services will be clarified in a future version. 96. The MPLS-TP control plane must support both E-LSP and L- LSP as specified in [RFC3270]. 97. The address spaces used in the management, control and data planes are independent. 98. The MPLS-TP control plane is based on the MPLS control plane for pseudowires, and more specifically, LDP is used for PW signaling. 99. Both single-segment and multi-segment PWs shall be supported by the MPLS-TP control plane. MPLS-TP shall use the definition of multi-segment PWs that is under development in the IETF independent from MPLS-TP. 100. The MPLS-TP control plane is based on the GMPLS control plane for MPLS-TP LSPs. More specifically, GMPLS RSVP-TE [RFC3473] and related extensions are used for LSP signaling, and GMPLS OSPF-TE [RFC5392] and ISIS-TE [RFC5316] are used for routing. 101. The MPLS-TP LSP control plane must allow for interoperation with the MPLS-TE LSP control plane. 102. The MPLS-TP control plane must ensure its own survivability and to enable it to recover gracefully from failures and degradations. These include graceful restart and hot redundant configurations. Andersson, et al Informational [Page 16] Internet-Draft draft-abfb-mpls-tp-control-plane-framework-01.txt July 13, 2009 103. The MPLS-TP control plane must support linear, ring and meshed protection schemes. 2.3. OAM Framework Derived Requirements The following additional requirements are based on [TP-OAM-REQ] and [TP-OAM]: 104. The MPLS-TP control plane must support the capability to enable/disable OAM functions as part of service establishment. 105. The MPLS-TP control plane must support the capability to enable/disable OAM functions after service establishment. In such cases, the customer must not perceive service degradation as a result of OAM enabling/disabling. 106. The MPLS-TP control plane must allow for the IP/MPLS and PW OAM protocols (e.g., LSP-Ping [RFC4379], MPLS-BFD [BFD-MPLS], VCCV [RFC5085] and VCCV-BFD [VCCV-BFD]). 107. The MPLS-TP control plane must allow for the ability to support experimental OAM functions. These functions must be disabled by default. 108. The MPLS-TP control plane must support the choice of which (if any) OAM function(s) to use and on which PW, LSP or Section to apply it(them) to. 109. The MPLS-TP control plane must provide a mechanism to support the localization of faults and the notification of appropriate nodes. Such notification should trigger corrective (recovery) actions. 110. The MPLS-TP control plane must allow service provider to be informed of a fault or defect affecting the service(s) it provides, even if the fault or defect is located outside of his domain. 111. Information exchange between various nodes involved in the MPLS-TP control plane should be reliable such that, for example, defects or faults are properly detected or that state changes are effectively known by the appropriate nodes. 112. The MPLS-TP control plane must provide functionality to control the verification of the continuity (checks) (CC) of a PW, LSP or Section. Andersson, et al Informational [Page 17] Internet-Draft draft-abfb-mpls-tp-control-plane-framework-01.txt July 13, 2009 113. The MPLS-TP control plane must provide functionality to control the verification of the connectivity (CV) of a PW, LSP or Section. 114. The MPLS-TP control plane may provide functionality to control the conduction of diagnostic tests on a PW, LSP or Section. 115. The MPLS-TP control plane must provide functionality to enable an End Point to discover the Intermediate (if any) and End Point(s) along a PW, LSP or Section, and more generally to trace (record) the route of a PW, LSP or Section. 116. The MPLS-TP control plane must provide functionality to enable an End Point of a PW, LSP or Section to instruct its associated End Point(s) to lock the PW, LSP or Section. Note that lock corresponds to an administrative status in which forwarding traffic on and from the PW, LSP or Section is disabled. (This requirement duplicates a requirement stated above but is listed again to maintain alignment with [TP-OAM].) 117. The MPLS-TP control plane must provide functionality to enable an Intermediate Point of a PW or LSP to report, to an End Point of that same PW or LSP, an external lock condition affecting that PW or LSP. 118. The MPLS-TP control plane must provide functionality to enable an Intermediate Point of a PW or LSP to report, to an End Point of that same PW or LSP, a fault or defect condition affecting that PW or LSP. 119. The MPLS-TP control plane must provide functionality to enable an End Point to report, to its associated End Point, a fault or defect condition that it detects on a PW, LSP or Section for which they are the End Points. 120. The MPLS-TP control plane must provide functionality to enable the propagation, across an MPLS-TP network, of information pertaining to a client defect of fault condition detected at an End Point of a PW or LSP, if the client layer mechanisms do not provide an alarm notification/propagation mechanism. 121. The MPLS-TP control plane must provide functionality to enable the control of quantification of packet loss ratio over a PW, LSP or Section. 122. The MPLS-TP control plane must provide functionality to report an expected one-way, and if appropriate, the two-way, delay of a PW, LSP or Section. Andersson, et al Informational [Page 18] Internet-Draft draft-abfb-mpls-tp-control-plane-framework-01.txt July 13, 2009 123. The MPLS-TP control plane must support the configuration of MEPs. a. The CC and CV functions operate between MEPs. b. All OAM packets coming to a MEP source are tunneled via label stacking, and therefore a MEP may only be present at an LSP's ingress and egress nodes (and never at an LSP's transit node). c. The CC and CV functions may serve as a trigger for protection switching, see requirement 45 above. d. This implies that LSP hierarchy must be used in cases where OAM is used to trigger recovery. 124. The MPLS-TP control plane must support the signaling of the transmission period and the ME identifier used in CC and CV. 2.4. Security Requirements There are no specific MPLS-TP control plane security requirements. The existing framework for MPLS and GMPLS security is documented on [MPLS-SEC] and that document applies equally to MPLS-TP. 3. Relationship of PWs and TE LSPs [Editor's note: This section (and the remainder of this document) is preliminary and will be edited/replaced in future versions.] TBD 4. TE LSPs [Editor's note: This section (and the remainder of this document) is preliminary and will be edited/replaced in future versions.] 4.1. General reuse of existing GMPLS control plane mechanisms As described in [RFC3945], Generalized MPLS (GMPLS) extends MPLS to support additional switching technologies. GMPLS is thus capable of controlling packet technologies. Most of the initial efforts on Generalized MPLS (GMPLS) have been related to delivering circuit connectivity. With the emergence of devices capable of multiple switching capabilities and the integrated control paradigm, there is a need to clarify the applicability of GMPLS to packet switching technologies. In particular, the formal definition of FAs and Andersson, et al Informational [Page 19] Internet-Draft draft-abfb-mpls-tp-control-plane-framework-01.txt July 13, 2009 hierarchy in [RFC4206] led to the definition of four regions for PSC (Packet Switching Capable) interfaces: PSC-1, PSC-2, PSC-3, and PSC-4. This document describes the GMPLS topics specifically related to MPLS-TP. 4.1.1. In-Band and Out-Of-Band The terms in-band and out-of-band typically refer to the relationship of the management and control planes relative to the data plane. The terms may be used to refer to the management plane independent of the control plane, or to both of them in concert. There are multiple uses of the terms in-band and out-of-band, and they may relate to a channel, a path or a network. Each of these can be used independently or in combination. Briefly, the terms are typically used as follows: o In-band This term is used to refer to cases where management and/or control plane traffic is sent using or embedded in the same communication channel used to transport the associated data. IP forwarded, MPLS packet, and Ethernet networks are all examples where control traffic is typically sent in-band with the data traffic. o Out-of-band, in-fiber This term is used to refer to cases where management and/or control plane traffic is sent using a different communication channel from the associated data traffic, and the control/management communication channel resides in the same fiber as the data traffic. Optical transport networks typically operate in an out-of-band in-fiber configuration. o Out-of-band, aligned topology This term is used to refer to the cases where management and/or control plane traffic is sent using a different communication channel from the associated data traffic, and the control/management communication must follow the same node-to- node path as the data traffic. Such topologies are usually supported using a parallel fiber or other configuration where multiple data channels are available and one is (dynamically) selected as the control channel. o Out-of-band, independent topology This term is used to refer to the cases where management and/or control plane traffic is sent using a different communication channel from the associated data traffic, and the control/management communication may follow a path that is completely independent of the data traffic. Such configurations don't preclude the use of in-fiber or aligned topology links, but alignment is not required. Andersson, et al Informational [Page 20] Internet-Draft draft-abfb-mpls-tp-control-plane-framework-01.txt July 13, 2009 In the context of MPLS-TP, requirement 4 can be met using out-of-band in-fiber or aligned topology types of control. Requirement 5 can only be met by using Out-of-band, independent topology. GMPLS routing and signaling can be used to support in-band and all of the out-of-band forms of control, see [RFC3945]. 4.1.2. Addressing MPLS-TP uses the IPv4 and IPv6 address families to identify MPLS-TP nodes by default for network management and signaling purposes. The separation of the control and management planes from the data plane allows each plane to be independently addressable. Each plane may use addresses that are not mutually reachable, e.g., it is likely that the data plane will not be able to reach an address from the management or control planes and vice versa. Each plane may also use a different address family. It is even possible to reuse addresses in each plane, but this is not recommended as it is likely lead to operational confusion. Unnumbered interfaces and links are also permitted and usage is at the discretion of the network operator. 4.2. Signaling In this section, we reference the existing MPLS and GMPLS signaling and routing mechanisms which can be used to support MPLS-TP LSPs. When controlling a packet-switched data-plane with GMPLS, the packets have an MPLS (see [RFC3032]) format, with the so-called "shim header" including a 20-bit label. Unlike MPLS, GMPLS uses the Generalized Label Object defined in [RFC3471] to signal such labels. In the current RSVP-TE signaling protocol, many objects make use of the Generalized Label. According to [RFC3471], a Generalized Label has the following format: "Generic MPLS labels and Frame Relay labels are encoded right justified aligned in 32 bits (4 octets)". This is primarily used in RESV messages to encode the downstream assigned label which shall be used on a link or FA of an LSP, using the LABEL object (class = 16). When the C-Type is set to 2, this LABEL object is carrying a Generalized Label encoded as defined in [RFC3471]. When a node wishes to restrict the set of labels possibly assigned by its downstream neighbor (for the LSP), it can use the LABEL_SET object in PATH messages: the Label Type must be set to "Generalized Label" (value=2) and the Sub-Channels must be such Generalized Labels. The SUGGESTED_LABEL, RECOVERY_LABEL and UPSTREAM_LABEL objects Andersson, et al Informational [Page 21] Internet-Draft draft-abfb-mpls-tp-control-plane-framework-01.txt July 13, 2009 (respectively, class = 129, 34, 35; C-Type = 2) of the PATH messages have an identical format to that of the Generalized Label Object. Similarly, the RECORD_ROUTE object of the PATH message can record the labels which are used along the LSP, using the label subobject TLV (type = 3). In this subobject, the C-type of the recorded label is copied (value is therefore 2 in the packet case), and the Label Object is copied into the appropriate field. The Generalized Label Request Object must be used in PATH messages (C-Type = 4) instead of the simple Label Request without range such as defined in [RFC3209] (C-Type = 1). In this object the Switching Type is then set to PSC-1, PSC-2, PSC-3 or PSC-4 (respectively values 1 to 4) according to the type of LSP being opened (see Section 3). The ACCEPTABLE_LABEL_SET object (Class= 130, C-Type = 1) of the PathErr message has an identical format to that of the LABEL_SET object of PATH messages. An MPLS-TP domain may be a switching point for an LSP that extends between client network islands. In this case, the MPLS- TP domain edge that connects to the respective client domain may have a static switching in the data plane done on the interface connecting to the respective client node. Alternatively, the LSP may be signaled between the client network and the MPLS-TP domain. There are two cases: (1) the client network connects via a GMPLS UNI to the MPLS-TP domain with knowledge of the remote MPLS-TP edge node and link that connects to the remote client node or there is some reachability information exchanged between the MPLS-TP domain and the client network via dynamic protocol, or (2) integrated model whereby the client network is an integrated part of the MPLS-TP domain, less likely option in some of the operation environments. 4.3. Routing The major extension in the context of routing PSC-LSPs within the GMPLS framework is the use of the various PSC-regions introduced by [RFC3945]. With the introduction of the hierarchy, formally specified in [RFC4206], it is necessary to use PSC-x as Switching Capability (SC) and therefore, the nesting process is modified with regards to the MPLS procedures. In particular, the policy chosen for announcing the SC associated with a Forwarding Adjacency has a significant impact. That is an MPLS-TP announced as an FA in a client network in an integrated model to support hierarchical MPLS-TP in MPLS-TP domain. Andersson, et al Informational [Page 22] Internet-Draft draft-abfb-mpls-tp-control-plane-framework-01.txt July 13, 2009 4.3.1. ISIS-TE/OSPF-TE routing in support of MPLS-TP The major extension in the context of routing PSC-LSPs within the GMPLS framework is the use of the various PSC-regions introduced by [RFC3945]. In MPLS, no hierarchy being formally defined, no limitations were applied on nesting packet LSPs within other packet LSPs. With the introduction of the hierarchy, formally specified in [RFC4206], it is necessary to use PSC-x as Switching Capability (SC) and therefore, the nesting process is modified with regards to the MPLS procedures. In particular, the policy chosen for announcing the SC associated with a Forwarding Adjacency has a significant impact. 4.3.1.1. ISIS-TE/OSPF-TE routing for MPLS-TP Per [RFC4203] for OSPF and [RFC5307] for IS-IS, the Interface Switching Capability Descriptor (ISCD) is a sub-TLV (of type 15) of the Link TLV, which is used to indicate the Switching Capability (or Capabilities) of an interface. Per [RFC4203], this TLV indicates encoding, MTU and bandwidth available at each priority level. The TLV also carries a Switching Capability field which indicates the switching hierarchy level: 1: Packet-Switch Capable-1 (PSC-1) 2: Packet-Switch Capable-2 (PSC-2) 3: Packet-Switch Capable-3 (PSC-3) 4: Packet-Switch Capable-4 (PSC-4) 4.3.1.2. Multiple Switching Capabilities To support interfaces that have more than one ISCD (see Section "Interface Switching Capability Descriptor" of [RFC4202]), the ISCD may occur more than once within a single routing protocol link description message. This allows a single packet TE-link or FA to be announced in multiple PSC regions, both as a PSC-1 and PSC-2 for instance. The "regular" packet TE-links (non-FAs) can also be configured to be used by one or several of the regions. A TE-link set to a single PSC- x region will be reserved for establishing PSC-x LSPs, whereas one set to multiple PSC-x, PSC-y and PSC-z regions will be shared by PSC LSPs of these switching types. When a TE-link or an FA is shared among regions, it is important for the nodes receiving traffic over this link/FA to have a single label- space shared across the regions. This is critical for the node to guarantee it will receive packets with different labels in different Andersson, et al Informational [Page 23] Internet-Draft draft-abfb-mpls-tp-control-plane-framework-01.txt July 13, 2009 packet regions, even when they arrive on the same interface. The fact that the label-space must be cross-region is independent from the fact that label-spaces may be per- interface, per-tunnel, per-upstream neighbor or per-platform. 4.3.1.3. Hierarchy [RFC4206] defines network regions based on switching capabilities. The hierarchy of regions is novel in GMPLS and this section intends to clarify the hierarchy for PSC nodes, and the use of the various PSC-regions. According to [RFC4206], there are four PSC regions which are hierarchically ordered in the following way: PSC-1 < PSC-2 < PSC-3 < PSC-4, that is PSC-1 is the smallest SC and PSC-4 is the largest SC. Let us consider two consecutive nodes of an LSP, such that the first node's SC is PSC-x and the second node's is PSC-y. The first node is said to be at the border of two packet regions, with regard to that LSP, if PSC-y is larger than PSC-x (i.e.: x < y ). Similarly, the second node is said to be at the border of two packet regions, with regard to that LSP, if x > y. According to [RFC4202], "a unidirectional LSP must have the same sets of SCs at both ends". Additionally, such an LSP will only be routed over TE-links and/or FAs which have (at least) that SC (since otherwise, the region crossing would trigger the setup of an FA-LSP, as described in [RFC4206]). This imposes that a PSC-x LSP be setup using only TE-links and/or FAs which include at least PSC-x. In the packet-switching context, this means that a PSC-x cannot directly use links/FAs which do not have a PSC-x set in their ISCD's Switching Capability Field. Therefore, if one wants to establish a PSC-x LSP across a PSC-y region, an FA- LSP must either be available or set-up. It may be announced in the PSC-x region routing instance (which may be the same as the PSC-y region routing instance) as a PSC-x TE-link. The SC associated with an FA is announced using the routing protocol's Interface Switching Capability Descriptor (ISCD) (see Section 3.1). For instance, if a PSC-1 LSP has to be setup across a PSC- 3 region, the region border node will first have to establish a PSC-3 LSP in which the PSC-1 LSP will be nested. The PSC-3 LSP may then be used to announce an FA in the PSC-1 routing protocol. In the four packet regions, the switching principles are the same, which means that a PSC node is most likely to have in fact all four PSC-1, PSC-2, PSC-3 and PSC-4 switching types. When using a packet LSP to nest other LSPs, the policy for deciding which PSCs to announce for the packet FAs and TE-links, and the policy for cross- region LSP triggering determine the type of interactions between the PSC-regions. This means there are in fact multiple ways of using the PSC regions. Andersson, et al Informational [Page 24] Internet-Draft draft-abfb-mpls-tp-control-plane-framework-01.txt July 13, 2009 4.3.2. TE link bundling 4.4. OAM, MEP (hierarchy) configuration & control Current MPLS LSP and PW OAM capabilities are not suitable for transport applications. Hence IETF has started work to define a comprehensive set of MPLS-TP OAM functions. Specific OAM requirements for MPLS-TP are documented in [draft-ietf-mpls-tp- oam-requirements]. In addition to the actual OAM requirements, it is also required that the control plane is able to configure and control OAM entities. This requirement is not yet addressed by the foreseen MPLS-TP control protocols (i.e, GMPLS for LSPs and T-LDP for PWs). To emphasize the importance of OAM establishment via the control plane it must be noted that for proper OAM; OAM messages and the actual normal traffic must be congruent: taking the same path and relying on the same forwarding decisions at intermediate nodes. Hence, it is desirable that OAM is setup together with the establishment of the data path (i.e., with the same signaling). This way OAM setup is bound to connection establishment signaling, avoiding two separate management/configuration steps (connection setup followed by OAM configuration) which would increases delay, processing and more importantly may be prune to misconfiguration errors. It must be noted that although the control plane is used to establish OAM entities, subsequently OAM is executed independently from the control plane. That is, OAM mechanisms are responsible for monitoring and initiating recovery actions (driving switches between primary and backup paths). GMPLS RSVP-TE based OAM configuration and control should be general to be applicable to a wide range of data plane technologies and OAM solution and not be limited to the MPLS technology and MPLS-TP OAM. On the other hand, GMPLS based OAM configuration must satisfy all MPLS-TP requirements. PW OAM establishment is FFS. 4.5. Traffic engineering and constraint-based path computation Same approach as MPLS. Specific algorithms out of scope. Similar to MPLS, but adds bidirectional and recovery path computation. Andersson, et al Informational [Page 25] Internet-Draft draft-abfb-mpls-tp-control-plane-framework-01.txt July 13, 2009 4.5.1. Relation to PCE Path Computation Element (PCE) may be used for path computation of a GMPLS LSP across domains and in a single domain. A Network Management System (NMS) may be used to trigger path computation for a GMPLS LSP and configure the cross-connects along the computed path. Alternatively, the path computation may be triggered by a network node via PCE Communication Protocol (PCECP) and the LSP signaled using GMPLS. 4.6. Applicability 4.7. Recovery 4.7.1. E2E, segment GMPLS defines recovery signaling for P2P LSPs in [RFC4872], RSVP-TE extensions in support for end-to-end GMPLS recovery, and [RFC4873], GMPLS segment recovery. GMPLS segment recovery provides a superset of the function in end-to-end recovery. The former can be viewed as a special case of segment recovery where there is a single recovery domain whose borders coincide with the ingress and egress of the LSP. All five of the protection types defined for recovery are supported in MPLS-TP. - 1+1 bidirectional protection for P2P LSPs - 1+1 unidirectional protection for P2MP LSPs - 1:n (including 1:1) protection with or without extra traffic - Rerouting without extra traffic (sometimes known as soft rerouting), including shared mesh restoration - Full LSP rerouting Recovery MUST be signaled using the mechanism defined in [RFC4872] and [RFC4873]. The use of Notify messages to trigger protection switching and recovery is not required in MPLS-TP as this function is expected to be supported via OAM. However, it's use is not precluded. The restoration priority for an MPLS-TP LSP is taken from the setup priority field in the SESSION_ATTRIBUTE object. The preemption priority for an MPLS-TP LSP is taken from the holding priority field in the SESSION_ATTRIBUTE object. Andersson, et al Informational [Page 26] Internet-Draft draft-abfb-mpls-tp-control-plane-framework-01.txt July 13, 2009 4.7.2. P2P, P2MP 4.7.3. Recovery Triggers The GMPLS control plane allows for management plane recovery triggers and directly supports control plane recovery triggers. Support for control plane recovery triggers is defined in [RFC4872] which refers to the triggers as "Recovery Commands". These commands can be used with both end-to-end and segment recovery, but are always controlled on an end-to-end basis. The recovery triggers/commands defined in [RFC4872] are: a. Lockout of recovery LSP b. Lockout of normal traffic c. Forced switch for normal traffic d. Requested switch for normal traffic e. Requested switch for recovery LSP Note that control plane triggers are typically invoked in response to a management plane request at the ingress. 4.8. Diffserv object usage in GMPLS (E-LSPs, L-LSPs) 4.9. Management plane support In MPLS-TP networks, like in any other technology, a control plane can be used to speed up and simplify provisioning and recovery actions in case of failures. The control plane needs, as does the data plane, to interact with the management plane in order to set- up/delete/modify LSPs and it also needs to be managed as any other network components by the Management Plane. [Editor's note: need to describe how the last sentence is supported via standard MIBs.] 4.9.1. Management Plane / Control Plane Ownership Transfer In networks where both control plane and management plane are provided, LSP provisioning can be bone either by control plane or management plane. As mentioned in the requirements section above, it must be possible to transfer, or handover, management plane created circuit under the control plane domain and vice-versa. [RFC5493] defines the specific requirements for an LSP ownership handover procedure. It must be possible for the control plane to notify, in reliable manner, the management plane about the status/result of either synchronous or asynchronous, with respect to the management plane, operation performed. Moreover it must be possible to monitor, via query or spontaneous notify, the status of the control plane object such as example the TE Link status, the available resources, etc. A mechanism must be made available by the control plane to the Andersson, et al Informational [Page 27] Internet-Draft draft-abfb-mpls-tp-control-plane-framework-01.txt July 13, 2009 management plane to log control plane LSP related operation, that is, it must be possible from the NMS to have a clear view of the life, (traffic hit, action performed, signaling etc.) of a given LSP. The LSP handover procedure for MPLS-TP LSPs is supported via [PC-SCP]. 4.10. CP reference points (E-NNI, I-NNI, UNI) 4.11. MPLS to MPLS-TP interworking - Leverage current MPLS and GMPLS development - Backward compatibility 5. Pseudo Wires [Editor's note: This section is preliminary and will be edited/replaced in future versions. There could be some differences (e.g., OAM interaction), and this section will focus on those differences.] MPLS Pseudo Wires, as defined in [RFC3985], provide for emulated services over an MPLS Packet Switched Network (PSN). There are several types of pseudowires: (1) Ethernet PWs providing for Ethernet port or Ethernet VLAN over MPLS [RFC4448], (2) HDLC/PPP Pseudowire providing for HDLC/PPP leased line transport of MPLS[RFC4618], (3) ATM PWs [RFC4816], (4) Frame Relay PWs [RFC4619], and (5) circulation Emulation PWs [RFC4553]. Today's transport networks based on PDH, WDM, or SONET/SDH provide transport for PDH or SONET (e.g., ATM over SONET or Packet PPP over SONET) client signals with no payload awareness. Implementing PW capability allows the use of an existing technology to substitute the TDM transport with Packet-aware transport, using well-defined pseudowire encapsulation methods for carrying various packet services over MPLS, and providing for potentially better bandwidth utilization. There are two types of pseudowires: (1) Single-Segment pseudowires (SS-PW), and (2) Multi-segment pseudowires (MS-PW). An MPLS-TP domain may transport a PW with endpoints within a client network transparently. Alternatively, an MPLS-TP edge node may be the Terminating PE (T-PE) for a PW, performing adaptation from the native attachment circuit technology (e.g. Ethernet 802.1q) to an MPLS PW for transport over an MPLS-TP domain, with a GMPLS LSP or a hierarchy of LSPs transporting the PW between the T-PEs. In this way, the PW is analogous to a transport channel in a TDM network and the LSP is equivalent to a container of multiple non-concatenated channels, albeit they are packet containers. The MPLS-TP domain may also contain Switching PEs (S-PEs) for a multi-segment PW whereby the T- Andersson, et al Informational [Page 28] Internet-Draft draft-abfb-mpls-tp-control-plane-framework-01.txt July 13, 2009 PEs may be at the edge of the MPLS-TP domain or in a client network. In this latter case, a T-PE in a client network is a T-PE performing the adaptation of the native service to MPLS and the MPLS-TP domain performs Pseudo-wire switching. SS-PW signaling control plane is based on LDP with specific procedures defined in [RFC4447]. [Segmented-PW] and [MS-PW] allow for static switching of multi-segment pseudowires in data and control plane and for dynamic routing and placement of an MS-PW whereby signaling is still based on Targeted LDP (T-LDP). The MPLS-TP domain shall use the same PW signaling protocols and procedures for placing SS-PWs and MS-PWs. This will leverage existing technology as well as facilitate interoperability with client networks with native attachment circuits or PW segment that is switched across the MPLS-TP domain. The same control protocol and procedures are reused as much as possible. However, when using PWs in MPLS-TP, a set of new requirements are defined which may require extensions of the existing control mechanisms. This section identifies areas where extensions are needed based on the PW Control Plane related requirements documented in [draft-ietf-mpls-tp-requirements]. The baseline requirement for extensions to support transport applications is that any new mechanisms and capabilities must be able to interoperate with existing IETF MPLS [RFC3031] and IETF PWE3 [RFC3985] control and data planes where appropriate. Hence, extensions of the PW Control Plane must be in-line with the procedures defined in [RFC4447]. For MPLS-TP, it is required that the data and control planes are both logically and physically separated. That is, the PW Control Plane must be able to operate out-of-band (OOB). This ensures that in the case of control plane failures the data plane is not affected and can continue to operate normally. This was not a design requirement for the current PW Control Plane. However, due to the PW concept, i.e., PWs are connecting logical entities ('forwarders'), and the operation of the PW control protocol, i.e., only edge PE nodes (T-PE, S-PE) take part in the signaling exchanges: moving T-LDP out-of-band seems to be, theoretically, a straightforward exercise. More precisely, if IP addressing is used in the MPLS-TP control plane then T-LDP addressing can be maintained, although all addresses will refer to control plane entities. Both, the PWid FEC and Generalized PWid FEC Elements can possibly be used in an OOB case as well (Detailed evaluation is FFS). The PW Label allocation and exchange mechanisms can be possibly reused unchanged (Detailed evaluation is FFS). Binding a PW to an LSP, or PW segments to LSPs can be left to networks elements acting as T-PEs and S-PEs or a control plane entity that may be the same one signaling the PW. If the control plane is physically separated from the forwarder, the control plane must be Andersson, et al Informational [Page 29] Internet-Draft draft-abfb-mpls-tp-control-plane-framework-01.txt July 13, 2009 able to program the forwarders with necessary information. For transport applications, it is mandatory that bidirectional traffic is following congruent paths. Today, each direction of a PW or a PW segment is bound to a unidirectional LSP that extends between two T-PEs, S-PEs, or a T-PE and an S-PE. The unidirectional LSPs in both directions are not required to following congruent paths, and therefore both directions of a PW may not follow congruent paths. The only requirement today is that a PW or a PW segment shares the same T-PEs in both directions, and same S-PEs in both directions. This poses a new requirement on the PW Control Plane, namely to ensure that both ends map the PW to the same transport path. When a bidirectional LSP is selected on one end to transport the PW, a mechanism is needed that signals to the remote end which LSP has been selected locally to transport the PW. This likely can be accomplished by adding a new TLV to PW signaling. This coincides with the gap identified for OOB support: a new mechanism may be needed to explicitly bind PWs to the supporting transport LSP. Alternatively, two unidirectional LSPs may be used to support the PW. However, to meet the congruency requirement, the LSPs must be placed so that they are forced to follow the same path (switches and links). This maybe accomplished by placing one unidirectional TE-LSP in one direction at one endpoint, and forcing the other endpoint to setup a TE-LSP with an ERO that has the nodes/links in the reverse order from the RRO seen in the path message of the LSP in the reverse direction. In this case, when one endpoint selects an LSP to bind the PW to, it must identify to the remote end which LSP to bind the other direction of the PW to. Transport applications require resource guarantees. In the case of transport LSPs, resource reservation mechanisms are provided via RSVP-TE and the use of DiffServ. If multiple PWs are multiplexed into the same transport LSP resources, contention may occur. However, local policy at PEs may ensure proper resource sharing among PWs mapped into a resource guaranteed LSP. On the other hand, it is limited if any guarantees can be provided to PWs if the LSPs used to support MPLS-TP PWs do not support resource guarantees. The PW control plane must be able to establish and configure all of the available features manageable for the PW, including protection and OAM entities and mechanisms. There is ongoing work on PW protection and MPLS-TP OAM. To summarize, the main areas identified for potential PW Control Plane extensions to support MPLS-TP are the following. Andersson, et al Informational [Page 30] Internet-Draft draft-abfb-mpls-tp-control-plane-framework-01.txt July 13, 2009 o Move PW Control Plane out-of-band o Explicit control of PW to LSP binding o PW QoS and congestion control o PW protection o PW OAM configuration and control 5.1. General reuse of existing PW control plane mechanisms 5.2. Signaling 5.3. Recovery (Redundancy) 6. Security Considerations [Editor's note: This section is preliminary and will be edited/replaced in future versions.] This document is a framework document and does not describe bits on the wire and have a very small impact on MPLS/GMPLS security issues. However it gives guidelines for future extension to existing MPLS and GMPLS protocols, it is understood that the documents that specify these extensions will address the security issues that relates to the extensions. The MPLS/GMPLS security framework [MPLS-SEC] is applicable to both this document and the documents that will be written as a result of the output of this document. 7. IANA Considerations There are no new IANA considerations introduced by this document. 8. Acknowledgments The authors would like to acknowledge the contributions of Yannick Brehon, Diego Caviglia, Nic Neate, and Dave Mcdysan to this work. Andersson, et al Informational [Page 31] Internet-Draft draft-abfb-mpls-tp-control-plane-framework-01.txt July 13, 2009 9. References 9.1. Normative References [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels," RFC 2119. [RFC3031] Rosen, E., Viswanathan, A., Callon, R., "Multiprotocol Label Switching Architecture", RFC 3031, January 2001. [RFC3032] Rosen, E., Tappan, D., Fedorkow, G., Rekhter, Y., Farinacci, D., Li, T., and Conta, A. "MPLS Label Stack Encoding", RFC 3032, January 2001. [RFC3209] Awduche, D., Berger, L., Gan, D., Li, T., Srinivasan, V., and G. Swallow, "RSVP-TE: Extensions to RSVP for LSP Tunnels", RFC 3209, December 2001. [RFC3471] Berger, L., "Generalized Multi-Protocol Label Switching (GMPLS) Signaling Functional Description", RFC 3471, January 2003. [RFC3473] Berger, L. Ed., "Generalized Multi-Protocol Label Switching (GMPLS) Signaling Resource ReserVation Protocol-Traffic Engineering (RSVP-TE) Extensions", January 2003. [RFC4202] Kompella, K. and Y. Rekhter, "Routing Extensions in Support of Generalized Multi-Protocol Label Switching(GMPLS)", RFC 4202, October 2005. [RFC4203] Kompella, K. and Y. Rekhter, "OSPF Extensions in Support of Generalized Multi-Protocol Label Switching (GMPLS)", RFC 4203, October 2005. [RFC4206] Kompella, K. and Y. Rekhter, "Label Switched Paths (LSP) Hierarchy with Generalized Multi-Protocol Label Switching (GMPLS) Traffic Engineering (TE)", RFC 4206, October 2005. [RFC4447] Martini, L., Ed., "Pseudowire Setup and Maintenance Using the Label Distribution Protocol (LDP)", RFC 4447, April 2006. [RFC4448] Martini, L., Ed., "Encapsulation Methods for Transport Ethernet over MPLS Network", RFC 4448, April 2006. Andersson, et al Informational [Page 32] Internet-Draft draft-abfb-mpls-tp-control-plane-framework-01.txt July 13, 2009 [RFC4872] Lang, J., Rekhter, Y., and Papadimitriou, D., "RSVP-TE Extensions in Support of End-to-End Generalized Multi- Protocol Label Switching (GMPLS) Recovery", RFC 4872, May 2007. [RFC4873] Berger, L., Bryskin, I., Papadimitriou, D., Farrel, A., "GMPLS Segment Recovery", RFC 4873, May 2007. [RFC4929] Andersson, L. and A. Farrel, "Change Process for Multiprotocol Label Switching (MPLS) and Generalized MPLS (GMPLS) Protocols and Procedures", BCP 129, RFC 4929, June 2007. [RFC5307] Kompella, K. and Rekhter, Y., "IS-IS Extensions in Support of Generalized Multi-Protocol Label Switching (GMPLS)", RFC 5307, October 2008. [RFC5316] Chen, M., Zhang, R., and Duan, X., "ISIS Extensions in Support of Inter-Autonomous System (AS) MPLS and GMPLS Traffic Engineering", RFC 5392, December 2008. [RFC5392] Chen, M., Zhang, R., and Duan, X., "OSPF Extensions in Support of Inter-Autonomous System (AS) MPLS and GMPLS Traffic Engineering", RFC 5392, January 2009. [TP-FWK] Bocci, M., Ed., Et al, "A Framework for MPLS in Transport Networks", work in Progress, draft-ietf-mpls-tp-framework-02, July 2009. [TP-OAM] Busi, I., Ed., Niven-Jenkins, B., Ed., "MPLS-TP OAM Framework and Overview", work in Progress, draft-busi-mpls-tp-oam-framework-02, March 2009. [TP-OAM-REQ] Vigoureux, M., Ward, D, and Betts, M., "Requirements for OAM in MPLS Transport Networks", draft-ietf-mpls-tp-oam-requirements-02, June 2009. [TP-SURVIVE] Sprecher, N., et al., "Multiprotocol Label Switching Transport Profile Survivability Framework", work in Progress, draft-sprecher-mpls-tp-survive-fwk-01.txt, February 2009. [TP-REQ] Niven-Jenkins, B., Et al, "MPLS-TP Requirements", work in Progress, draft-ietf-mpls-tp-requirements-09, June 2009. Andersson, et al Informational [Page 33] Internet-Draft draft-abfb-mpls-tp-control-plane-framework-01.txt July 13, 2009 9.2. Informative References [BFD-MPLS] Aggarwal, R., Kompella, K., Nadeau, T., and G. Swallow, "BFD For MPLS LSPs", draft-ietf-bfd-mpls-07 (work in progress), June 2008. [ITU.G8080.2006] International Telecommunications Union, "Architecture for the automatically switched optical network (ASON)", ITU- T Recommendation G.8080, June 2006. [ITU.G8080.2008] International Telecommunications Union, "Architecture for the automatically switched optical network (ASON) Amendment 1", ITU-T Recommendation G.8080 Amendment 1, March 2008. [MPLS-SEC] Fang, L., et al, "Security Framework for MPLS and GMPLS Networks", work in Progress, draft-ietf-mpls-mpls-and-gmpls-security-framework-05.txt, March 2009. [PC-SCP] Caviglia, D, et al, "RSVP-TE Signaling Extension For The Conversion Between Permanent Connections And Soft Permanent Connections In A GMPLS Enabled Transport Network.", draft-ietf-ccamp-pc-spc-rsvpte-ext-02.txt, work in progress, October 2008. [Segmented-PW] Martini, L., Nadeau, T., and Duckett M., " Segmented Pseaudowire", work in Progress, draft-ietf-pwe3-segmented-pw-12.txt, June 2009. [RFC3945] Mannie, E., "Generalized Multi-Protocol Label Switching (GMPLS) Architecture", RFC 3945, October 2004. [RFC3985] Bryant, S. and P. Pate, "Pseudo Wire Emulation Edge- to-Edge (PWE3) Architecture", RFC 3985, March 2005. [RFC3270] Le Faucheur, F., et al, "Multi-Protocol Label Switching (MPLS) Support of Differentiated Services", RFC3270, May 2002. [RFC4139] Papadimitriou, D., et al, "Requirements for Generalized MPLS (GMPLS) Signaling Usage and Extensions for Automatically Switched Optical Network (ASON)", RFC4139, July 2005. Andersson, et al Informational [Page 34] Internet-Draft draft-abfb-mpls-tp-control-plane-framework-01.txt July 13, 2009 [RFC4208] Swallow, G., Drake, J., Ishimatsu, H., and Rekhter, Y., "Generalized Multi-Protocol Label Switching (GMPLS) User-Network Interface (UNI): Resource ReserVation Protocol-Traffic Engineering (RSVP-TE) Support for the Overlay Model", RFC 4206, October 2005. [RFC4258] Brungard, D., et al, "Requirements for Generalized Multi-Protocol Label Switching (GMPLS) Routing for the Automatically Switched Optical Network (ASON)", RFC4258, November 2005. [RFC4379] Kompella, K. and G. Swallow, "Detecting Multi-Protocol Label Switched (MPLS) Data Plane Failures", RFC 4379, February 2006. [RFC4427] Mannie, E., Papadimitriou, D., "Recovery (Protection and Restoration) Terminology for Generalized Multi-Protocol Label Switching (GMPLS)", RFC4427, March 2006. [RFC4553] Vainshtein, A., Ed., and Stein, YJ., Ed.,"Structure- Agnostic Time Division Multiplexing (TDM) over Packet (SAToP)", RFC 4553, June 2006. [RFC4618] Martini, L., Rosen, E., Heron, G., and Malis, A., "Encapsulation Methods for Transport of PPP/High- Level Data Link Control (HDLC) over MPLS Networks", RFC 4618, September 2006. [RFC4619] Martini, L., Ed., Kawa, C., Ed., and Malis, A., Ed., "Encapsulation Methods for Transport of Frame Relay over Multiprotocol Label Switching (MPLS) Networks", September 2006. [RFC4816] Malis, A., Martini, L., Brayley, J., and Walsh, T., "Pseudowire Emulation Edge-to-Edge (PWE3) Asynchronous Transfer Mode (ATM) Transparent Cell Transport Service", RFC 4816, February 2007. [RFC5085] Nadeau, T. and C. Pignataro, "Pseudowire Virtual Circuit Connectivity Verification (VCCV): A Control Channel for Pseudowires", RFC 5085, December 2007. [RFC5493] Caviglia, D., et al, "Requirements for the Conversion between Permanent Connections and Switched Connections in a Generalized Multiprotocol Label Switching (GMPLS) Network", RFC 5493, April 2009. Andersson, et al Informational [Page 35] Internet-Draft draft-abfb-mpls-tp-control-plane-framework-01.txt July 13, 2009 [MS-PW] Bocci, M., and Bryant, B., "An Architecture for Multi-Segment Pseudowire Emulation Edge-to-Edge", work in Progress, draft-ietf-pwe3-ms-pw-arch-06.txt, February 2009. [VCCV-BFD] Nadeau, T. and C. Pignataro, "Bidirectional Forwarding Detection (BFD) for the Pseudowire Virtual Circuit Connectivity Verification (VCCV)", draft-ietf-pwe3-vccv-bfd-06 (work in progress), July 2009. 10. Authors' Addresses Loa Andersson (editor) Redback Networks Phone: +46 8 632 77 14 Email: loa@pi.nu Lou Berger (editor) LabN Consulting, L.L.C. Phone: +1-301-468-9228 Email: lberger@labn.net Luyuan Fang (editor) Cisco Systems, Inc. 300 Beaver Brook Road Boxborough, MA 01719 USA Email: lufang@cisco.com Nabil Bitar (editor) Verizon, 40 Sylvan Rd., Waltham, MA 02451 Email: nabil.n.bitar@verizon.com Attila Takacs Ericsson 1. Laborc u. Budapest, HUNGARY 1037 Email: attila.takacs@ericsson.com Martin Vigoureux Alcatel-Lucent Email: martin.vigoureux@alcatel-lucent.fr Andersson, et al Informational [Page 36] Generated on: Mon Jul 13 13:10:02 EDT 2009