Network Working Group E. Mannie, Ed. Request for Comments: 4427 Perceval Category: Informational D. Papadimitriou, Ed. Alcatel March 2006 Recovery (Protection and Restoration) Terminology for Generalized Multi-Protocol Label Switching (GMPLS) Status of This Memo This memo provides information for the Internet community. It does not specify an Internet standard of any kind. Distribution of this memo is unlimited. Copyright Notice Copyright (C) The Internet Society (2006). Abstract This document defines a common terminology for Generalized Multi- Protocol Label Switching (GMPLS)-based recovery mechanisms (i.e., protection and restoration). The terminology is independent of the underlying transport technologies covered by GMPLS. Mannie & Papadimitriou Informational [Page 1] RFC 4427 GMPLS Recovery Terminology March 2006 Table of Contents 1. Introduction ....................................................3 2. Contributors ....................................................4 3. Conventions Used in this Document ...............................5 4. Recovery Terminology Common to Protection and Restoration .......5 4.1. Working and Recovery LSP/Span ..............................6 4.2. Traffic Types ..............................................6 4.3. LSP/Span Protection and Restoration ........................6 4.4. Recovery Scope .............................................7 4.5. Recovery Domain ............................................8 4.6. Recovery Types .............................................8 4.7. Bridge Types ..............................................10 4.8. Selector Types ............................................10 4.9. Recovery GMPLS Nodes ......................................11 4.10. Switch-over Mechanism ....................................11 4.11. Reversion operations .....................................11 4.12. Failure Reporting ........................................12 4.13. External commands ........................................12 4.14. Unidirectional versus Bi-Directional Recovery Switching ..13 4.15. Full versus Partial Span Recovery Switching ..............14 4.16. Recovery Schemes Related Time and Durations ..............14 4.17. Impairment ...............................................15 4.18. Recovery Ratio ...........................................15 4.19. Hitless Protection Switch-over ...........................15 4.20. Network Survivability ....................................15 4.21. Survivable Network .......................................16 4.22. Escalation ...............................................16 5. Recovery Phases ................................................16 5.1. Entities Involved During Recovery .........................17 6. Protection Schemes .............................................17 6.1. 1+1 Protection ............................................18 6.2. 1:N (N >= 1) Protection ...................................18 6.3. M:N (M, N > 1, N >= M) Protection .........................18 6.4. Notes on Protection Schemes ...............................19 7. Restoration Schemes ............................................19 7.1. Pre-Planned LSP Restoration ...............................19 7.1.1. Shared-Mesh Restoration ............................19 7.2. LSP Restoration ...........................................20 7.2.1. Hard LSP Restoration ...............................20 7.2.2. Soft LSP Restoration ...............................20 8. Security Considerations ........................................20 9. References .....................................................20 9.1. Normative References ......................................20 9.2. Informative References ....................................20 10. Acknowledgements ..............................................21 Mannie & Papadimitriou Informational [Page 2] RFC 4427 GMPLS Recovery Terminology March 2006 1. Introduction This document defines a common terminology for Generalized Multi- Protocol Label Switching (GMPLS)-based recovery mechanisms (i.e., protection and restoration). The terminology proposed in this document is independent of the underlying transport technologies and borrows from the G.808.1 ITU-T Recommendation [G.808.1] and from the G.841 ITU-T Recommendation [G.841]. The restoration terminology and concepts have been gathered from numerous sources including IETF documents. In the context of this document, the term "recovery" denotes both protection and restoration. The specific terms "protection" and "restoration" will only be used when differentiation is required. This document focuses on the terminology for the recovery of Label Switched Paths (LSPs) controlled by a GMPLS control plane. The proposed terminology applies to end-to-end, segment, and span (i.e., link) recovery. Note that the terminology for recovery of the control plane itself is not in the scope of this document. Protection and restoration of switched LSPs under tight time constraints is a challenging problem. This is particularly relevant to optical networks that consist of Time Division Multiplex (TDM) and/or all-optical (photonic) cross-connects referred to as GMPLS nodes (or simply nodes, or even sometimes "Label Switching Routers, or LSRs") connected in a general topology [RFC3945]. Recovery typically involves the activation of a recovery (or alternate) LSP when a failure is encountered in the working LSP. A working or recovery LSP is characterized by an ingress interface, an egress interface, and a set of intermediate nodes and spans through which the LSP is routed. The working and recovery LSPs are typically resource disjoint (e.g., node and/or span disjoint). This ensures that a single failure will not affect both the working and recovery LSPs. A bi-directional span between neighboring nodes is usually realized as a pair of unidirectional spans. Therefore, the end-to-end path for a bi-directional LSP consists of a series of bi-directional segments (i.e., Sub-Network Connections, or SNCs, in the ITU-T terminology) between the source and destination nodes, traversing intermediate nodes. Mannie & Papadimitriou Informational [Page 3] RFC 4427 GMPLS Recovery Terminology March 2006 2. Contributors This document is the result of a joint effort by the CCAMP Working Group Protection and Restoration design team. The following are the authors that contributed to the present document: Deborah Brungard (AT&T) Rm. D1-3C22 - 200 S. Laurel Ave. Middletown, NJ 07748, USA EMail: dbrungard@att.com Sudheer Dharanikota EMail: sudheer@ieee.org Jonathan P. Lang (Sonos) 506 Chapala Street Santa Barbara, CA 93101, USA EMail: jplang@ieee.org Guangzhi Li (AT&T) 180 Park Avenue, Florham Park, NJ 07932, USA EMail: gli@research.att.com Eric Mannie Perceval Rue Tenbosch, 9 1000 Brussels Belgium Phone: +32-2-6409194 EMail: eric.mannie@perceval.net Dimitri Papadimitriou (Alcatel) Francis Wellesplein, 1 B-2018 Antwerpen, Belgium EMail: dimitri.papadimitriou@alcatel.be Mannie & Papadimitriou Informational [Page 4] RFC 4427 GMPLS Recovery Terminology March 2006 Bala Rajagopalan Microsoft India Development Center Hyderabad, India EMail: balar@microsoft.com Yakov Rekhter (Juniper) 1194 N. Mathilda Avenue Sunnyvale, CA 94089, USA EMail: yakov@juniper.net 3. 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]. 4. Recovery Terminology Common to Protection and Restoration This section defines the following general terms common to both protection and restoration (i.e., recovery). In addition, most of these terms apply to end-to-end, segment, and span LSP recovery. Note that span recovery does not protect the nodes at each end of the span, otherwise end-to-end or segment LSP recovery should be used. The terminology and the definitions were originally taken from [G.808.1]. However, for generalization, the following language, which is not directly related to recovery, has been adapted to GMPLS and the common IETF terminology: An LSP is used as a generic term to designate either an SNC (Sub- Network Connection) or an NC (Network Connection) in ITU-T terminology. The ITU-T uses the term transport entity to designate either a link, an SNC, or an NC. The term "Traffic" is used instead of "Traffic Signal". The term protection or restoration "scheme" is used instead of protection or restoration "architecture". The reader is invited to read [G.841] and [G.808.1] for references to SDH protection and Generic Protection Switching terminology, respectively. Note that restoration is not in the scope of [G.808.1]. Mannie & Papadimitriou Informational [Page 5] RFC 4427 GMPLS Recovery Terminology March 2006 4.1. Working and Recovery LSP/Span A working LSP/span is an LSP/span transporting "normal" user traffic. A recovery LSP/span is an LSP/span used to transport "normal" user traffic when the working LSP/span fails. Additionally, the recovery LSP/span may transport "extra" user traffic (i.e., pre-emptable traffic) when normal traffic is carried over the working LSP/span. 4.2. Traffic Types The different types of traffic that can be transported over an LSP/span, in the context of this document, are defined hereafter: A. Normal traffic: User traffic that may be protected by two alternative LSPs/spans (the working and recovery LSPs/spans). B. Extra traffic: User traffic carried over recovery resources (e.g., a recovery LSP/span) when these resources are not being used for the recovery of normal traffic (i.e., when the recovery resources are in standby mode). When the recovery resources are required to recover normal traffic from the failed working LSP/span, the extra traffic is pre- empted. Extra traffic is not protected by definition, but may be restored. Moreover, extra traffic does not need to commence or be terminated at the ends of the LSPs/spans that it uses. C. Null traffic: Traffic carried over the recovery LSP/span if it is not used to carry normal or extra traffic. Null traffic can be any kind of traffic that conforms to the signal structure of the specific layer, and it is ignored (not selected) at the egress of the recovery LSP/span. 4.3. LSP/Span Protection and Restoration The following subtle distinction is generally made between the terms "protection" and "restoration", even though these terms are often used interchangeably [RFC3386]. The distinction between protection and restoration is made based on the resource allocation done during the recovery LSP/span establishment. The distinction between different types of restoration is made based on the level of route computation, signaling, and resource allocation during the restoration LSP/span establishment. Mannie & Papadimitriou Informational [Page 6] RFC 4427 GMPLS Recovery Terminology March 2006 A. LSP/Span Protection LSP/span protection denotes the paradigm whereby one or more dedicated protection LSP(s)/span(s) is/are fully established to protect one or more working LSP(s)/span(s). For a protection LSP, this implies that route computation took place, that the LSP was fully signaled all the way, and that its resources were fully selected (i.e., allocated) and cross-connected between the ingress and egress nodes. For a protection span, this implies that the span has been selected and reserved for protection. Indeed, it means that no signaling takes place to establish the protection LSP/span when a failure occurs. However, various other kinds of signaling may take place between the ingress and egress nodes for fault notification, to synchronize their use of the protection LSP/span, for reversion, etc. B. LSP/Span Restoration LSP/span restoration denotes the paradigm whereby some restoration resources may be pre-computed, signaled, and selected a priori, but not cross-connected to restore a working LSP/span. The complete establishment of the restoration LSP/span occurs only after a failure of the working LSP/span, and requires some additional signaling. Both protection and restoration require signaling. Signaling to establish the recovery resources and signaling associated with the use of the recovery LSP(s)/span(s) are needed. 4.4. Recovery Scope Recovery can be applied at various levels throughout the network. An LSP may be subject to local (span), segment, and/or end-to-end recovery. Local (span) recovery refers to the recovery of an LSP over a link between two nodes. End-to-end recovery refers to the recovery of an entire LSP from its source (ingress node end-point) to its destination (egress node end- point). Segment recovery refers to the recovery over a portion of the network of a segment LSP (i.e., an SNC in the ITU-T terminology) of an end- to-end LSP. Such recovery protects against span and/or node failure Mannie & Papadimitriou Informational [Page 7] RFC 4427 GMPLS Recovery Terminology March 2006 over a particular portion of the network that is traversed by an end-to-end LSP. 4.5. Recovery Domain A recovery domain is defined as a set of nodes and spans, over which one or more recovery schemes are provided. A recovery domain served by one single recovery scheme is referred to as a "single recovery domain", while a recovery domain served by multiple recovery schemes is referred to as a "multi recovery domain". The recovery operation is contained within the recovery domain. A GMPLS recovery domain must be entirely contained within a GMPLS domain. A GMPLS domain (defined as a set of nodes and spans controlled by GMPLS) may contain multiple recovery domains. 4.6. Recovery Types The different recovery types can be classified depending on the number of recovery LSPs/spans that are protecting a given number of working LSPs/spans. The definitions given hereafter are from the point of view of a working LSP/span that needs to be protected by a recovery scheme. A. 1+1 type: dedicated protection One dedicated protection LSP/span protects exactly one working LSP/span, and the normal traffic is permanently duplicated at the ingress node on both the working and protection LSPs/spans. No extra traffic can be carried over the protection LSP/span. This type is applicable to LSP/span protection, but not to LSP/span restoration. B. 0:1 type: unprotected No specific recovery LSP/span protects the working LSP/span. However, the working LSP/span can potentially be restored through any alternate available route/span, with or without any pre-computed restoration route. Note that no resources are pre-established for this recovery type. This type is applicable to LSP/span restoration, but not to LSP/span protection. Span restoration can be achieved, for instance, by moving all the LSPs transported over a failed span to a dynamically selected span. Mannie & Papadimitriou Informational [Page 8] RFC 4427 GMPLS Recovery Terminology March 2006 C. 1:1 type: dedicated recovery with extra traffic One specific recovery LSP/span protects exactly one specific working LSP/span, but the normal traffic is transmitted over only one LSP (working or recovery) at a time. Extra traffic can be transported using the recovery LSP/span resources. This type is applicable to LSP/span protection and LSP restoration, but not to span restoration. D. 1:N (N > 1) type: shared recovery with extra traffic A specific recovery LSP/span is dedicated to the protection of up to N working LSPs/spans. The set of working LSPs/spans is explicitly identified. Extra traffic can be transported over the recovery LSP/span. All these LSPs/spans must start and end at the same nodes. Sometimes, the working LSPs/spans are assumed to be resource disjoint in the network so that they do not share any failure probability, but this is not mandatory. Obviously, if more than one working LSP/span in the set of N are affected by some failure(s) at the same time, the traffic on only one of these failed LSPs/spans may be recovered over the recovery LSP/span. Note that N can be arbitrarily large (i.e., infinite). The choice of N is a policy decision. This type is applicable to LSP/span protection and LSP restoration, but not to span restoration. Note: a shared recovery where each recovery resource can be shared by a maximum of X LSPs/spans is not defined as a recovery type but as a recovery scheme. The choice of X is a network resource management policy decision. E. M:N (M, N > 1, N >= M) type: A set of M specific recovery LSPs/spans protects a set of up to N specific working LSPs/spans. The two sets are explicitly identified. Extra traffic can be transported over the M recovery LSPs/spans when available. All the LSPs/spans must start and end at the same nodes. Sometimes, the working LSPs/spans are assumed to be resource disjoint in the network so that they do not share any failure probability, but this is not mandatory. Obviously, if several working LSPs/spans in the set of N are concurrently affected by some failure(s), the traffic on only M of these failed LSPs/spans may be recovered. Note that N can be arbitrarily large (i.e., infinite). The choice of N and M is a policy decision. Mannie & Papadimitriou Informational [Page 9] RFC 4427 GMPLS Recovery Terminology March 2006 This type is applicable to LSP/span protection and LSP restoration, but not to span restoration. 4.7. Bridge Types A bridge is the function that connects the normal traffic and extra traffic to the working and recovery LSP/span. A. Permanent bridge Under a 1+1 type, the bridge connects the normal traffic to both the working and protection LSPs/spans. This type of bridge is not applicable to restoration types. There is, of course, no extra traffic connected to the recovery LSP/span. B. Broadcast bridge For 1:N and M:N types, the bridge permanently connects the normal traffic to the working LSP/span. In the event of recovery switching, the normal traffic is additionally connected to the recovery LSP/span. Extra traffic is either not connected or connected to the recovery LSP/span. C. Selector bridge For 1:N and M:N types, the bridge connects the normal traffic to either the working or the recovery LSP/span. Extra traffic is either not connected or connected to the recovery LSP/span. 4.8. Selector Types A selector is the function that extracts the normal traffic from either the working or the recovery LSP/span. Extra traffic is either extracted from the recovery LSP/span, or is not extracted. A. Selective selector Is a selector that extracts the normal traffic from either the working LSP/span output or the recovery LSP/span output. B. Merging selector For 1:N and M:N protection types, the selector permanently extracts the normal traffic from both the working and recovery LSP/span outputs. This alternative works only in combination with a selector bridge. Mannie & Papadimitriou Informational [Page 10] RFC 4427 GMPLS Recovery Terminology March 2006 4.9. Recovery GMPLS Nodes This section defines the GMPLS nodes involved during recovery. A. Ingress GMPLS node of an end-to-end LSP/segment LSP/span The ingress node of an end-to-end LSP/segment LSP/span is where the normal traffic may be bridged to the recovery end-to-end LSP/segment LSP/span. Also known as source node in the ITU-T terminology. B. Egress GMPLS node of an end-to-end LSP/segment LSP/span The egress node of an end-to-end LSP/segment LSP/span is where the normal traffic may be selected from either the working or the recovery end-to-end LSP/segment LSP/span. Also known as sink node in the ITU-T terminology. C. Intermediate GMPLS node of an end-to-end LSP/segment LSP A node along either the working or recovery end-to-end LSP/segment LSP route between the corresponding ingress and egress nodes. Also known as intermediate node in the ITU-T terminology. 4.10. Switch-over Mechanism A switch-over is an action that can be performed at both the bridge and the selector. This action is as follows: A. For the selector: The action of selecting normal traffic from the recovery LSP/span rather than from the working LSP/span. B. For the bridge: In case of permanent connection to the working LSP/span, the action of connecting or disconnecting the normal traffic to or from the recovery LSP/span. In case of non-permanent connection to the working LSP/span, the action of connecting the normal traffic to the recovery LSP/span. 4.11. Reversion operations A revertive recovery operation refers to a recovery switching operation, where the traffic returns to (or remains on) the working LSP/span when the switch-over requests are terminated (i.e., when the working LSP/span has recovered from the failure). Mannie & Papadimitriou Informational [Page 11] RFC 4427 GMPLS Recovery Terminology March 2006 Therefore, a non-revertive recovery switching operation is when the traffic does not return to the working LSP/span when the switch-over requests are terminated. 4.12. Failure Reporting This section gives (for information) several signal types commonly used in transport planes to report a failure condition. Note that fault reporting may require additional signaling mechanisms. A. Signal Degrade (SD): a signal indicating that the associated data has degraded. B. Signal Fail (SF): a signal indicating that the associated data has failed. C. Signal Degrade Group (SDG): a signal indicating that the associated group data has degraded. D. Signal Fail Group (SFG): a signal indicating that the associated group has failed. Note: SDG and SFG definitions are under discussion at the ITU-T. 4.13. External commands This section defines several external commands, typically issued by an operator through the Network Management System (NMS)/Element Management System (EMS), that can be used to influence or command the recovery schemes. A. Lockout of recovery LSP/span: A configuration action, initiated externally, that results in the recovery LSP/span being temporarily unavailable to transport traffic (either normal or extra traffic). B. Lockout of normal traffic: A configuration action, initiated externally, that results in the normal traffic being temporarily not allowed to be routed over its recovery LSP/span. Note that in this case extra-traffic is still allowed on the recovery LSP/span. Mannie & Papadimitriou Informational [Page 12] RFC 4427 GMPLS Recovery Terminology March 2006 C. Freeze: A configuration action, initiated externally, that prevents any switch-over action from being taken, and, as such, freezes the current state. D. Forced switch-over for normal traffic: A switch-over action, initiated externally, that switches normal traffic to the recovery LSP/span, unless an equal or higher priority switch-over command is in effect. E. Manual switch-over for normal traffic: A switch-over action, initiated externally, that switches normal traffic to the recovery LSP/span, unless a fault condition exists on other LSPs/spans (including the recovery LSP/span) or an equal or higher priority switch-over command is in effect. F. Manual switch-over for recovery LSP/span: A switch-over action, initiated externally, that switches normal traffic to the working LSP/span, unless a fault condition exists on the working LSP/span or an equal or higher priority switch-over command is in effect. G. Clear: An action, initiated externally, that clears the active external command. 4.14. Unidirectional versus Bi-Directional Recovery Switching A. Unidirectional recovery switching: A recovery switching mode in which, for a unidirectional fault (i.e., a fault affecting only one direction of transmission), only the normal traffic transported in the affected direction (of the LSP or span) is switched to the recovery LSP/span. B. Bi-directional recovery switching: A recovery switching mode in which, for a unidirectional fault, the normal traffic in both directions (of the LSP or span), including the affected direction and the unaffected direction, are switched to the recovery LSP/span. Mannie & Papadimitriou Informational [Page 13] RFC 4427 GMPLS Recovery Terminology March 2006 4.15. Full versus Partial Span Recovery Switching Bulk LSP recovery is initiated upon reception of either span failure notification or bulk failure notification of the S LSPs carried by this span. In either case, the corresponding recovery switching actions are performed at the LSP level, such that the ratio between the number of recovery switching messages and the number of recovered LSP (in one given direction) is minimized. If this ratio equals 1, one refers to full span recovery; otherwise, if this ratio is greater than 1, one refers to partial span recovery. A. Full Span Recovery All the S LSP carried over a given span are recovered under span failure condition. Full span recovery is also referred to as "bulk recovery". B. Partial Span Recovery Only a subset s of the S LSP carried over a given span is recovered under span failure condition. Both selection criteria of the entities belonging to this subset, and the decision concerning the recovery of the remaining (S - s) LSP, are based on local policy. 4.16. Recovery Schemes Related Time and Durations This section gives several typical timing definitions that are of importance for recovery schemes. A. Detection time: The time between the occurrence of the fault or degradation and its detection. Note that this is a rather theoretical time because, in practice, this is difficult to measure. B. Correlation time: The time between the detection of the fault or degradation and the reporting of the signal fail or degrade. This time is typically used in correlating related failures or degradations. C. Notification time: The time between the reporting of the signal fail or degrade and the reception of the indication of this event by the entities that decide on the recovery switching operation(s). Mannie & Papadimitriou Informational [Page 14] RFC 4427 GMPLS Recovery Terminology March 2006 D. Recovery Switching time: The time between the initialization of the recovery switching operation and the moment the normal traffic is selected from the recovery LSP/span. E. Total Recovery time: The total recovery time is defined as the sum of the detection, the correlation, the notification, and the recovery switching time. F. Wait To Restore time: A period of time that must elapse after a recovered fault before an LSP/span can be used again to transport the normal traffic and/or to select the normal traffic from. Note: the hold-off time is defined as the time between the reporting of signal fail or degrade, and the initialization of the recovery switching operation. This is useful when multiple layers of recovery are being used. 4.17. Impairment A defect or performance degradation, which may lead to SF or SD trigger. 4.18. Recovery Ratio The quotient of the actual recovery bandwidth divided by the traffic bandwidth that is intended to be protected. 4.19. Hitless Protection Switch-over Protection switch-over, which does not cause data loss, data duplication, data disorder, or bit errors upon recovery switching action. 4.20. Network Survivability The set of capabilities that allows a network to restore affected traffic in the event of a failure. The degree of survivability is determined by the network's capability to survive single and multiple failures. Mannie & Papadimitriou Informational [Page 15] RFC 4427 GMPLS Recovery Terminology March 2006 4.21. Survivable Network A network that is capable of restoring traffic in the event of a failure. 4.22. Escalation A network survivability action caused by the impossibility of the survivability function in lower layers. 5. Recovery Phases It is commonly accepted that recovery implies that the following generic operations need to be performed when an LSP/span or a node failure occurs: - Phase 1: Failure Detection The action of detecting the impairment (defect of performance degradation) as a defect condition and the consequential activation of SF or SD trigger to the control plane (through internal interface with the transport plane). Thus, failure detection (which should occur at the transport layer closest to the failure) is the only phase that cannot be achieved by the control plane alone. - Phase 2: Failure Localization (and Isolation) Failure localization provides, to the deciding entity, information about the location (and thus the identity) of the transport plane entity that causes the LSP(s)/span(s) failure. The deciding entity can then make an accurate decision to achieve finer grained recovery switching action(s). - Phase 3: Failure Notification Failure notification phase is used 1) to inform intermediate nodes that LSP(s)/span(s) failure has occurred and has been detected and 2) to inform the recovery deciding entities (which can correspond to any intermediate or end-point of the failed LSP/span) that the corresponding LSP/span is not available. - Phase 4: Recovery (Protection or Restoration) See Section 4.3. - Phase 5: Reversion (Normalization) See Section 4.11. Mannie & Papadimitriou Informational [Page 16] RFC 4427 GMPLS Recovery Terminology March 2006 The combination of Failure Detection and Failure Localization and Notification is referred to as Fault Management. 5.1. Entities Involved During Recovery The entities involved during the recovery operations can be defined as follows; these entities are parts of ingress, egress, and intermediate nodes, as defined previously: A. Detecting Entity (Failure Detection): An entity that detects a failure or group of failures; thus providing a non-correlated list of failures. B. Reporting Entity (Failure Correlation and Notification): An entity that can make an intelligent decision on fault correlation and report the failure to the deciding entity. Fault reporting can be automatically performed by the deciding entity detecting the failure. C. Deciding Entity (part of the failure recovery decision process): An entity that makes the recovery decision or selects the recovery resources. This entity communicates the decision to the impacted LSPs/spans with the recovery actions to be performed. D. Recovering Entity (part of the failure recovery activation process): An entity that participates in the recovery of the LSPs/spans. The process of moving failed LSPs from a failed (working) span to a protection span must be initiated by one of the nodes that terminates the span, e.g., A or B. The deciding (and recovering) entity is referred to as the "master", while the other node is called the "slave" and corresponds to a recovering only entity. Note: The determination of the master and the slave may be based on configured information or protocol-specific requirements. 6. Protection Schemes This section clarifies the multiple possible protection schemes and the specific terminology for the protection. Mannie & Papadimitriou Informational [Page 17] RFC 4427 GMPLS Recovery Terminology March 2006 6.1. 1+1 Protection 1+1 protection has one working LSP/span, one protection LSP/span, and a permanent bridge. At the ingress node, the normal traffic is permanently bridged to both the working and protection LSP/span. At the egress node, the normal traffic is selected from the better of the two LSPs/spans. Due to the permanent bridging, the 1+1 protection does not allow an unprotected extra traffic signal to be provided. 6.2. 1:N (N >= 1) Protection 1:N protection has N working LSPs/spans that carry normal traffic and 1 protection LSP/span that may carry extra-traffic. At the ingress, the normal traffic is either permanently connected to its working LSP/span and may be connected to the protection LSP/span (case of broadcast bridge), or is connected to either its working LSP/span or the protection LSP/span (case of selector bridge). At the egress node, the normal traffic is selected from either its working or protection LSP/span. Unprotected extra traffic can be transported over the protection LSP/span whenever the protection LSP/span is not used to carry a normal traffic. 6.3. M:N (M, N > 1, N >= M) Protection M:N protection has N working LSPs/spans carrying normal traffic and M protection LSP/span that may carry extra-traffic. At the ingress, the normal traffic is either permanently connected to its working LSP/span and may be connected to one of the protection LSPs/spans (case of broadcast bridge), or is connected to either its working LSP/span or one of the protection LSPs/spans (case of selector bridge). At the egress node, the normal traffic is selected from either its working or one of the protection LSP/span. Unprotected extra traffic can be transported over the M protection LSP/span whenever the protection LSPs/spans is not used to carry a normal traffic. Mannie & Papadimitriou Informational [Page 18] RFC 4427 GMPLS Recovery Terminology March 2006 6.4. Notes on Protection Schemes All protection types are either uni- or bi-directional; obviously, the latter applies only to bi-directional LSPs/spans and requires coordination between the ingress and egress node during protection switching. All protection types except 1+1 unidirectional protection switching require a communication channel between the ingress and the egress node. In the GMPLS context, span protection refers to the full or partial span recovery of the LSPs carried over that span (see Section 4.15). 7. Restoration Schemes This section clarifies the multiple possible restoration schemes and the specific terminology for the restoration. 7.1. Pre-Planned LSP Restoration Also referred to as pre-planned LSP re-routing. Before failure detection and/or notification, one or more restoration LSPs are instantiated between the same ingress-egress node pair as the working LSP. Note that the restoration resources must be pre-computed, must be signaled, and may be selected a priori, but may not cross- connected. Thus, the restoration LSP is not able to carry any extra-traffic. The complete establishment of the restoration LSP (i.e., activation) occurs only after failure detection and/or notification of the working LSP and requires some additional restoration signaling. Therefore, this mechanism protects against working LSP failure(s) but requires activation of the restoration LSP after failure occurrence. After the ingress node has activated the restoration LSP, the latter can carry the normal traffic. Note: when each working LSP is recoverable by exactly one restoration LSP, one refers also to 1:1 (pre-planned) re-routing without extra- traffic. 7.1.1. Shared-Mesh Restoration "Shared-mesh" restoration is defined as a particular case of pre- planned LSP re-routing that reduces the restoration resource requirements by allowing multiple restoration LSPs (initiated from distinct ingress nodes) to share common resources (including links and nodes.) Mannie & Papadimitriou Informational [Page 19] RFC 4427 GMPLS Recovery Terminology March 2006 7.2. LSP Restoration Also referred to as LSP re-routing. The ingress node switches the normal traffic to an alternate LSP that is signaled and fully established (i.e., cross-connected) after failure detection and/or notification. The alternate LSP path may be computed after failure detection and/or notification. In this case, one also refers to "Full LSP Re-routing." The alternate LSP is signaled from the ingress node and may reuse the intermediate node's resources of the working LSP under failure condition (and may also include additional intermediate nodes.) 7.2.1. Hard LSP Restoration Also referred to as hard LSP re-routing. A re-routing operation where the LSP is released before the full establishment of an alternate LSP (i.e., break-before-make). 7.2.2. Soft LSP Restoration Also referred to as soft LSP re-routing. A re-routing operation where the LSP is released after the full establishment of an alternate LSP (i.e., make-before-break). 8. Security Considerations Security considerations are detailed in [RFC4428] and [RFC4426]. 9. References 9.1. Normative References [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, March 1997. 9.2. Informative References [RFC3386] Lai, W. and D. McDysan, "Network Hierarchy and Multilayer Survivability", RFC 3386, November 2002. [RFC3945] Mannie, E., "Generalized Multi-Protocol Label Switching (GMPLS) Architecture", RFC 3945, October 2004. [RFC4426] Lang, J., Rajagopalan B., and D.Papadimitriou, Editors, "Generalized Multiprotocol Label Switching (GMPLS) Recovery Functional Specification", RFC 4426, March 2006. Mannie & Papadimitriou Informational [Page 20] RFC 4427 GMPLS Recovery Terminology March 2006 [RFC4428] Papadimitriou D. and E.Mannie, Editors, "Analysis of Generalized Multi-Protocol Label Switching (GMPLS)-based Recovery Mechanisms (including Protection and Restoration)", RFC 4428, March 2006. For information on the availability of the following documents, please see http://www.itu.int [G.808.1] ITU-T, "Generic Protection Switching - Linear trail and subnetwork protection," Recommendation G.808.1, December 2003. [G.841] ITU-T, "Types and Characteristics of SDH Network Protection Architectures," Recommendation G.841, October 1998. 10. Acknowledgements Many thanks to Adrian Farrel for having thoroughly review this document. Editors' Addresses Eric Mannie Perceval Rue Tenbosch, 9 1000 Brussels Belgium Phone: +32-2-6409194 EMail: eric.mannie@perceval.net Dimitri Papadimitriou Alcatel Francis Wellesplein, 1 B-2018 Antwerpen, Belgium Phone: +32 3 240-8491 EMail: dimitri.papadimitriou@alcatel.be Mannie & Papadimitriou Informational [Page 21] RFC 4427 GMPLS Recovery Terminology March 2006 Full Copyright Statement Copyright (C) The Internet Society (2006). This document is subject to the rights, licenses and restrictions contained in BCP 78, and except as set forth therein, the authors retain all their rights. This document and the information contained herein are provided on an "AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS OR IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY AND THE INTERNET ENGINEERING TASK FORCE DISCLAIM 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. Intellectual Property The IETF takes no position regarding the validity or scope of any Intellectual Property Rights or other rights that might be claimed to pertain to the implementation or use of the technology described in this document or the extent to which any license under such rights might or might not be available; nor does it represent that it has made any independent effort to identify any such rights. Information on the procedures with respect to rights in RFC documents can be found in BCP 78 and BCP 79. Copies of IPR disclosures made to the IETF Secretariat and any assurances of licenses to be made available, or the result of an attempt made to obtain a general license or permission for the use of such proprietary rights by implementers or users of this specification can be obtained from the IETF on-line IPR repository at http://www.ietf.org/ipr. The IETF invites any interested party to bring to its attention any copyrights, patents or patent applications, or other proprietary rights that may cover technology that may be required to implement this standard. Please address the information to the IETF at ietf-ipr@ietf.org. Acknowledgement Funding for the RFC Editor function is provided by the IETF Administrative Support Activity (IASA). Mannie & Papadimitriou Informational [Page 22]