CCAMP Working Group D. Papadimitriou (Alcatel) Internet Draft Z. Lin (New York City Transit) Category: Informational J. Drake (Calient) J. Ash (ATT) Expiration Date: February 2004 A. Farrel (Old Dog Consulting) L. Ong (Ciena) August 2003 Requirements for Generalized MPLS (GMPLS) Signaling Usage and Extensions for Automatically Switched Optical Network (ASON) draft-ietf-ccamp-gmpls-ason-reqts-02.txt Status of this Memo This document is an Internet-Draft and is in full conformance with all provisions of Section 10 of RFC-2026. 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/ietf/1id-abstracts.txt The list of Internet-Draft Shadow Directories can be accessed at http://www.ietf.org/shadow.html. 1. Abstract The Generalized MPLS (GMPLS) suite of protocols has been defined to control different switching technologies as well as different applications. These include support for requesting TDM connections including SONET/SDH and Optical Transport Networks (OTNs). This document concentrates on the signaling aspects of the GMPLS suite of protocols. It identifies the features to be covered by the GMPLS signaling protocol to support the capabilities of an Automatically Switched Optical Network (ASON). This document provides a problem statement and additional requirements on the GMPLS signaling protocol to support the ASON functionality. D.Papadimitriou et al. - Expires February 2004 1 draft-ietf-ccamp-gmpls-ason-reqts-02.txt August 2003 2. 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 RFC-2119. 3. Introduction The GMPLS suite of protocol specifications provides support for controlling different switching technologies as well as different applications. These include support for requesting TDM connections including SONET/SDH (see ANSI T1.105 and ITU-T G.707, respectively) as well as Optical Transport Networks (see ITU-T G.709). In addition, there are certain capabilities that are needed to support Automatically Switched Optical Networks control planes (their architecture is defined in [ITU-T G.8080]). These include generic capabilities such as call and connection separation, and more specific capabilities such as support of soft permanent connections. This document concentrates on the signaling aspects of the GMPLS suite of protocols. It discusses the functional requirements that lead to additional and backward compatible extensions to GMPLS signaling (see [RFC 3471] and [RFC 3473]) to support the capabilities as specified in the above referenced document. A description of backward compatibility considerations is provided in Section 5. A terminology section is provided in the Appendix. Problem Statement: The Automatically Switched Optical Network (ASON) architecture describes the application of an automated control plane for supporting both call and connection management services (for a detailed description see [ITU-T G.8080]). The ASON control plane specification is meant to be applicable to different transport technologies (e.g., SDH/SONET, OTN) in various networking environments (e.g., inter-carrier, intra-carrier). Also, the ASON model distinguishes reference points (representing points of protocol information exchange) defined (1) between an administrative domain and a user, (2) between administrative domains and, (3) between areas of the same administrative domain and when needed between control components (or simply controllers) within areas. A full description of the ASON terms and relationship between ASON model and GMPLS protocol suite may be found in [IPO-ASON]. This document describes the use of GMPLS signaling (in particular, [RFC 3471] and [RFC 3473]) to provide call and connection management (see [ITU-T G.7713]). The following functionality is expected to be supported and to be backward compatible with the GMPLS protocol suite as currently defined by the IETF: (a) soft permanent connection capability (b) call and connection separation D.Papadimitriou et al. - Expires February 2004 2 draft-ietf-ccamp-gmpls-ason-reqts-02.txt August 2003 (c) call segments (d) extended restart capabilities during control plane failures (e) extended label usage (f) crankback capability (g) additional error cases. 4. Requirements for Extending Applicability of GMPLS to ASON The applicability statements regarding how the GMPLS suite of protocols may be applied to the ASON architecture can be found in [IPO-ASON] and [IPO-REQS]. The former includes a summary of the ASON functions as well as a detailed discussion of the applicability of the GMPLS protocol suite. The next sections detail the requirements concerning the functions including: - Support for soft permanent connection capability - Support for call and connection separation - Support for call segments - Support for extended restart capabilities during control plane failures - Support for extended label usage - Support for crankback capability - Support for additional error cases Also, the support of these functions is strictly independent and must be agnostic of any user-to-network interface and therefore not constrained or restricted by its implementation specifics (see [ITU- T G.8080] and [ITU-T G.7713]). In support of the G.8080 end-to-end call model across different signaling domains, end-to-end signaling should be facilitated regardless of the administrative boundaries and protocols within the network. The resulting requirement being that there needs to be a clear mapping of signaling requests between GMPLS control domains and non-GMPLS control domains. This document provides signalling requirements for G.8080 distributed call and connection management based on GMPLS, within a GMPLS based control domain and between GMPLS based control domains. It does not restrict use of other protocols within a control domain. Interworking aspects, including mapping of non-GMPLS protocol signaling requests and support of non- GMPLS address formats, is outside the scope of this document. Compatibility aspects of non-GMPLS systems (nodes) within a GMPLS control domain i.e. the support of GMPLS systems and other systems which utilize other signaling protocols or some which may not support any signaling protocols is described. For instance, Section 4.5 'Support for Extended Label Usage' covers the requirements when a non-GMPLS capable sub-network is introduced or when nodes do not support any signaling protocols. 4.1 Support for Soft Permanent Connection (SPC) Capability D.Papadimitriou et al. - Expires February 2004 3 draft-ietf-ccamp-gmpls-ason-reqts-02.txt August 2003 An SPC is a combination of a permanent connection at the source user-to-network side, a permanent connection at the destination user-to-network side, and a switched connection within the network. An Element Management System (EMS) or a Network Management System (NMS) typically initiates the establishment of the switched connection by communicating with the node that initiates the switched connection (also known as the ingress node). The latter then sets the connection using the distributed GMPLS signaling protocol. For the SPC, the communication method between the EMS/NMS and the ingress node is beyond the scope of this document (so it is for any other function described in this document). The end-to-end connection is thus created by associating the incoming interface of the ingress node with the switched connection within the network, and the outgoing interface of the switched connection terminating network node (also referred to as egress node). An SPC connection is illustrated in the following Figure, which shows user's node A connected to a provider's node B via link #1, user's node Z connected to a provider's node Y via link #3, and an abstract link #2 connecting provider's node B and node Y. --- --- --- --- | A |--1--| B |-----2-//------| Y |--3--| Z | --- --- --- --- In this instance, the connection on link #1 and link #3 are both provisioned (permanent connections that may be simple links). In contrast, the connection over link #2 is set up using the distributed control plane. Thus the SPC is composed of the splicing of link #1, #2 and #3. Thus, to support the capability to request an SPC connection: - The GMPLS signaling protocol must be capable of supporting the ability to indicate the outgoing link and label information used when setting up the destination provisioned connection. - In addition, due to the inter-domain applicability of ASON networks, the GMPLS signaling protocol should also support indication of the service level requested for the SPC. In the case where an SPC spans multiple domains, indication of both source and destination endpoints controlling the SPC request may be needed. These may be done via the source and destination signaling controller addresses. Note that the association at the ingress node between the permanent connection and the switched connection is an implementation matter under the control of the EMS/NMS and is not within the scope of the signaling protocol. It is, therefore, outside the scope of this document. 4.2 Support for Call and Connection Separation D.Papadimitriou et al. - Expires February 2004 4 draft-ietf-ccamp-gmpls-ason-reqts-02.txt August 2003 A call may be simply described as "An association between endpoints that supports an instance of a service" [ITU-T G.8080]. Thus, it can be considered as a service provided between two end-points, where several calls may exist between them. Multiple connections may be associated to each call. The call concept provides an abstract relationship between two users, where this relationship describes (or verifies) to what extent the users are willing to offer (or accept) service to each other. Therefore, a call does not provide the actual connectivity for transmitting user traffic, but only builds a relationship by which subsequent connections may be made. A property of a call is to contain zero, one or multiple connections. Within the same call, connections may be of different types and each connection may exist independently of other connections, i.e., each connection is setup and released with separate Path/Resv messages. For example, a call may contain a set of basic connections and virtually concatenated connections (see [GMPLS-SONET] for corresponding connection signaling extensions). The concept of the call allows for a better flexibility in how end- points set up connections and how networks offer services to users. In essence, a call allows: - Support for virtual concatenation where each connection can travel on different diverse paths - An upgrade strategy for control plane operations, where a call control component (service provisioning) may be separate from the actual nodes hosting the connections (where the connection control component may reside) - Identification of the call initiator (with both network call controller as well as destination user) prior to connection, which may result in decreasing contention during resource reservation - General treatment of multiple connections which may be associated for several purposes; for example a pair of working and recovery connections may belong to the same call. To support the introduction of the call concept, GMPLS signaling should include a call identification mechanism and allow for end-to- end call capability exchange. For instance, a feasible structure for the call identifier (to guarantee global uniqueness) may concatenate a globally unique fixed ID (e.g., may be composed of country code, carrier code) with an operator specific ID (where the operator specific ID may be composed of a unique access point code - such as source node address - and a local identifier). Other formats shall also be possible depending on the call identification conventions between parties involved in the call setup process. 4.3 Support for Call Segments As described in [ITU-T G.8080], call segmentation may be applied when a call crosses several administrative domains. As such, an end- D.Papadimitriou et al. - Expires February 2004 5 draft-ietf-ccamp-gmpls-ason-reqts-02.txt August 2003 to-end call may consist of multiple call segments, when the call traverses multiple administrative domains. Each call segment can have one or more associated connections and the number of connections associated with each call segment may not be the same for a given end-to-end call. The initiating caller interacts with a called party by means of one or more intermediate call controllers located at the network edge between administrative domains (i.e., inter-domain reference point) and in particular at the user-to-network reference point. Their functions are defined by the policies associated by interactions between the administrative domain boundaries and between users and the network. This capability allows for independent (policy based) choices of signaling, concatenation, data plane protection and control plane driven recovery paradigms in different administrative domains. 4.4 Support for Extended Restart Capabilities Various types of failures may occur affecting the ASON control plane. Requirements placed on the control plane failure recovery by [ITU-T G.8080] include: - Any control plane failure must not result in releasing established calls and connections. - Upon recovery from a control plane failure, the recovered node must have the ability to recover the status of the calls and connections established before failure occurrence. - Upon recovery from a control plane failure, the recovered node must have the ability to recover the connectivity information of its neighbors. - Upon recovery from a control plane failure, calls and connections in the process of being established (i.e. pending call/connection setup requests) should be released or continued (with setup). - Upon recovery from a control plane failure, calls and connections in the process of being released must be released. - Upon recovery from a control plane failure, a call must have the ability to re-synchronize with its associated connections. 4.5 Support for Extended Label Usage Labels are defined in GMPLS (see [RFC 3471]) to provide information on the resources used on link local basis for a particular connection. The labels may range from specifying a particular timeslot, a particular wavelength to a particular port/fiber. In the ASON context, the value of a label MAY not be consistently the same across a link. For example, the figure below illustrates the case where two GMPLS capable nodes (A and Z) are interconnected across two non-GMPLS capable nodes (B and C), where these nodes are all SONET/SDH nodes providing, e.g., a VC-4 service. D.Papadimitriou et al. - Expires February 2004 6 draft-ietf-ccamp-gmpls-ason-reqts-02.txt August 2003 ----- ----- | | --- --- | | | A |---| B |---| C |---| Z | | | --- --- | | ----- ----- Labels have an associated implicit imposed structure based on [GMPLS-SONET] and [GMPLS-OTN]. Thus, once the local label is exchanged with its neighboring control plane node, the structure of the local label MAY not be significant to the neighbor node since the association between the local and the remote label may not necessarily be the same. This issue does not present a problem in simple point-to-point connections between two control plane-enabled nodes where the timeslots are mapped 1:1 across the interface. However, once a non-GMPLS capable sub-network is introduced between these nodes (as in the above figure, where the sub-network provides re-arrangement capability for the timeslots) label scoping MAY become an issue. In this context, there is an implicit assumption that the data plane connections between the GMPLS capable edges already exist prior to any connection request. For instance, node A's outgoing VC-4's timeslot #1 (with SUKLM label=[1,0,0,0,0]) as defined in [GMPLS- SONET]) may be mapped onto node B's outgoing VC-4's timeslot #6 (label=[6,0,0,0,0]) may be mapped onto node C's outgoing VC-4's timeslot #4 (label=[4,0,0,0,0]). Thus by the time node Z receives the request from node A with label=[1,0,0,0,0], the node Z's local label and the timeslot no longer corresponds to the received label and timeslot information. As such, to support this capability, a label association mechanism has to be used by the control plane node to map the received (remote) label into a locally significant label. The information necessary to allow mapping from received label value to a locally significant label value may be derived in several ways including: - Manual provisioning of the label association - Discovery of the label association Either method may be used. In case of dynamic association, this implies that the discovery mechanism operates at the timeslot/label level before the connection request is processed at the ingress node. Note that in the case where two nodes are directly connected, no association is required. In particular, for directly connected TDM interfaces no mapping function (at all) is required due to the implicit label structure (see [GMPLS-SONET] and [GMPLS-OTN]). In such instances, the label association function provides a one-to-one mapping of the received to local label values. 4.6 Support for Crankback D.Papadimitriou et al. - Expires February 2004 7 draft-ietf-ccamp-gmpls-ason-reqts-02.txt August 2003 Crankback has been identified as an important requirement for ASON networks. It allows a connection setup request to be retried on an alternate path that detours around a blocked link or node upon a setup failure, for instance, because a link or a node along the selected path has insufficient resources. Crankback mechanisms may also be applied during connection recovery by indicating the location of the failed link or node. This would significantly improve the successful recovery ratio for failed connections, especially in situations where a large number of setup requests are simultaneously triggered. The following mechanisms are assumed during crankback signaling (see also [GMPLS-CRANK]): - the blocking resource (link or node) must be identified and returned in the error response message towards the repair node (that may or may not be the ingress node); it is also assumed that this process will occur within a limited period of time - the computation (from the repair node) of an alternate path around the blocking link or node satisfying the initial connection constraints - the re-initiation of the connection setup request from the repair node (i.e. the node that has intercepted and processed the error response message) The following properties are expected for crankback signaling (see [GMPLS-CRANK]): - Error information persistence: the entity that computes the alternate (re-routing) path should store the identifiers of the blocking resources as indicated in the error message until the connection is successfully established or until the node abandons rerouting attempts. Since crankback may happen more than once while establishing a specific connection, the history of all experienced blockages for this connection should be maintained (at least until the routing protocol updates the state of this information) to perform an accurate path computation avoiding all blockages. - Rerouting attempts limitation: to prevent an endless repetition of connection setup attempts (using crankback information), the number of retries should be strictly limited. The maximum number of crankback rerouting attempts allowed can be limited per connection, per node, per area or even per administrative domain. - When the number of retries at a particular node or area is exceeded, the node currently handling the failure reports the error message upstream to the next repair node where further rerouting attempts may be performed. It is important that the crankback information provided indicates that re-routing through this node will not succeed. - When the maximum number of retries for a specific connection has been exceeded, the repair node handling the current failure should send an error message upstream indicating "Maximum D.Papadimitriou et al. - Expires February 2004 8 draft-ietf-ccamp-gmpls-ason-reqts-02.txt August 2003 number of re-routings exceeded". This error message will be sent back to the ingress node with no further rerouting attempts. Then, the ingress node may choose to retry the connection setup according to local policy but also re-use its original path or compute a path that avoids the blocking resources. Note: after several retries, a given repair point may be unable to compute a path to the destination node that avoids all of the blockages. In this case, it must pass the error message upstream to the next repair point. 4.7 Support for Additional Error Cases To support the ASON network, the following additional category of error cases are defined: - Errors associated with basic call and soft permanent connection support. For example, these may include incorrect assignment of IDs for the Call or an invalid interface ID for the soft permanent connection. - Errors associated with policy failure during processing of the new call and soft permanent connection capabilities. These may include unauthorized request for the particular capability. - Errors associated with incorrect specification of the service level. 5. Backward Compatibility As noted above, any extensions to the GMPLS signaling protocol in support of the requirements described in this document must be backward compatible. Backward compatibility means that in a network of nodes, some of which support GMPLS signaling extensions to facilitate the functions described in this document, and some of which do not, it must be possible to set up conventional connections (as described by [RFC 3473]) between any arbitrary pair of nodes and traversing any arbitrary set of nodes. Further, the use of any GMPLS signaling extensions to set up calls or connections that support the functions described in this document must not perturb existing conventional connections. Additionally, when transit nodes, that do not need to participate in the new functions described in this document, lie on the path of a call or connection, the GMPLS signaling extensions must be such that those transit nodes are able to participate in the establishment of the call or connection by passing the setup information onwards, unmodified. Lastly, when a transit or egress node is called upon to support a function described in this document, but does not, the GMPLS signaling extensions must be such that they can be rejected by pre- D.Papadimitriou et al. - Expires February 2004 9 draft-ietf-ccamp-gmpls-ason-reqts-02.txt August 2003 existing GMPLS signaling mechanisms in a way that is not detrimental to the network as a whole. 6. Security Considerations Per [ITU-T G.8080], a connection cannot be established until the associated call has been set up. Also, policy and authentication procedures are applied prior to the establishment of the call (and can then also be restricted to connection establishment in the context of this call). This document introduces no new security requirements to GMPLS signaling (see [RFC3471]). 7. Acknowledgements The authors would like to thank Deborah Brungard, Nic Larkin, Osama Aboul-Magd and Dimitrios Pendarakis for their comments and contributions to the previous version of this document. 8. References 8.1 Normative References [RFC-2026] S.Bradner, "The Internet Standards Process -- Revision 3", BCP 9, RFC 2026, October 1996. [RFC-2119] S.Bradner, "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, March 1997. [RFC-3209] D.Awduche et al., "RSVP-TE: Extensions to RSVP for LSP Tunnels," RFC 3209, December 2001. [RFC-3471] L.Berger (Editor) et al., "Generalized Multi- Protocol Label Switching (GMPLS) - Signaling Functional Description," RFC 3471, January 2003. [RFC-3473] L.Berger (Editor) et al., "Generalized Multi-Protocol Label Switching (GMPLS) Signaling - Resource ReserVation Protocol-Traffic Engineering (RSVP-TE) Extensions," RFC 3473, January 2003. [ITUT G.8080] ITU-T Rec. G.8080/Y.1304, "Architecture for the Automatically Switched Optical Network (ASON)," November 2001 (and Revision, January 2003). [GMPLS-CRANK] A.Farrel (Editor), "Crankback Routing Extensions for MPLS Signaling," Work in Progress, draft-iwata-mpls- crankback-06.txt, June 2003. [GMPLS-SONET] E.Mannie and D.Papadimitriou (Editors), "GMPLS Extensions for SONET and SDH Control, Work in D.Papadimitriou et al. - Expires February 2004 10 draft-ietf-ccamp-gmpls-ason-reqts-02.txt August 2003 Progress," draft-ietf-ccamp-gmpls-sonet-sdh-08.txt, February 2003. [GMPLS-OTN] D.Papadimitriou (Editor), "GMPLS Signaling Extensions for G.709 Optical Transport Networks Control," Work in progress, draft-ietf-ccamp-gmpls-g709-04.txt, May 2003. 8.2 Informative References [IPO-ASON] Aboul-Magd (Editor) et al., "Automatic Switched Optical Network (ASON) Architecture and Its Related Protocols," Work in progress, draft-ietf-ipo-ason- 02.txt, March 2002. [IPO-REQS] Y.Xue (Editor) et al., "Optical Network Service Requirements," Work in progress, draft-ietf-ipo- carrier-requirements-05.txt. [ITUT G.7713] ITU-T Rec. G.7713/Y.1304, "Distributed Call and Connection Management," November 2001. 9. Author's Addresses Dimitri Papadimitriou (Alcatel) Francis Wellesplein 1, B-2018 Antwerpen, Belgium Email: dimitri.papadimitriou@alcatel.be Zhi-Wei Lin (New York City Transit) 2 Broadway, Room C3.25 New York, NY 10004, USA Email: zhiwlin@nyct.com John Drake (Calient) 5853 Rue Ferrari, San Jose, CA 95138, USA Email: jdrake@calient.net Adrian Farrel (Old Dog Consulting) Email: adrian@olddog.co.uk Gerald R. Ash (ATT) AT&T Labs, Room MT D5-2A01 200 Laurel Avenue Middletown, NJ 07748, USA Email: gash@att.com Lyndon Ong (Ciena) 5965 Silver Creek Valley Road San Jose, CA 95138, USA Email: lyong@ciena.com D.Papadimitriou et al. - Expires February 2004 11 draft-ietf-ccamp-gmpls-ason-reqts-02.txt August 2003 Appendix - Terminology This document makes use of the following terms: Administrative domain: See Recommendation G.805. Call: association between endpoints that supports an instance of a service. Connection: concatenation of link connections and sub-network connections that allows the transport of user information between the ingress and egress points of a sub-network. Control plane: performs the call control and connection control functions. Through signaling, the control plane sets up and releases connections, and may restore a connection in case of a failure. (Control) Domain: represents a collection of entities that are grouped for a particular purpose. G.8080 applies this G.805 recommendation concept (that defines two particular forms, the administrative domain and the management domain) to the control plane in the form of a control domain. The entities that are grouped in a control domain are components of the control plane. External NNI: interfaces are located between protocol controllers between control domains. Internal NNI: interfaces are located between protocol controllers within control domains. Link: See Recommendation G.805. Management plane: performs management functions for the Transport Plane, the control plane and the system as a whole. It also provides coordination between all the planes. The following management functional areas are performed in the management plane: performance, fault, configuration, accounting and security management Management domain: See Recommendation G.805. Transport plane: provides bi-directional or unidirectional transfer of user information, from one location to another. It can also provide transfer of some control and network management information. The Transport Plane is layered; it is equivalent to the Transport Network defined in G.805. D.Papadimitriou et al. - Expires February 2004 12 draft-ietf-ccamp-gmpls-ason-reqts-02.txt August 2003 Full Copyright Statement "Copyright (C) The Internet Society (2003). All Rights Reserved. This document and translations of it may be copied and furnished to others, and derivative works that comment on or otherwise explain it or assist in its implementation may be prepared, copied, published and distributed, in whole or in part, without restriction of any kind, provided that the above copyright notice and this paragraph are included on all such copies and derivative works. However, this document itself may not be modified in any way, such as by removing the copyright notice or references to the Internet Society or other Internet organizations, except as needed for the purpose of developing Internet standards in which case the procedures for copyrights defined in the Internet Standards process must be followed, or as required to translate it into languages other than English. The limited permissions granted above are perpetual and will not be revoked by the Internet Society or its successors or assigns. This document and the information contained herein is provided on an "AS IS" basis and THE INTERNET SOCIETY AND THE INTERNET ENGINEERING TASK FORCE DISCLAIMS ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. D.Papadimitriou et al. - Expires February 2004 13