L2VPN Working Group Pranjal Kumar Dutta Florin Balus Internet Draft Alcatel-Lucent Intended status: Standard Expires: December 16, 2011 Olen Stokes Extreme Networks Geraldine Calvignac France Telecom June 23, 2011 LDP Extensions for Optimized MAC Address Withdrawal in H-VPLS draft-ietf-l2vpn-vpls-ldp-mac-opt-04.txt Status of this Memo This Internet-Draft is submitted in full conformance with the provisions of 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/ietf/1id-abstracts.txt The list of Internet-Draft Shadow Directories can be accessed at http://www.ietf.org/shadow.html This Internet-Draft will expire on December 23, 2011. Copyright Notice Copyright (c) 2011 IETF Trust and the persons identified as the document authors. All rights reserved. Dutta, et. al. Expires December 23, 2011 [Page 1] Internet-Draft Optimized MAC Withdrawal in H-VPLS June 23, 2011 This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (http://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License. Abstract [RFC4762] describes a mechanism to remove or unlearn MAC addresses that have been dynamically learned in a VPLS Instance for faster convergence on topology change. The procedure also removes MAC addresses in the VPLS that do not require relearning due to such topology change. This document defines an enhancement to the MAC Address Withdrawal procedure with empty MAC List [RFC4762], which enables a Provider Edge(PE) device to remove only the MAC addresses that need to be relearned. Additional extensions to [RFC4762] MAC Withdrawal procedures are specified to provide optimized MAC flushing for the PBB-VPLS specified in [PBB-VPLS Model]. Table of Contents 1.1. Conventions used in this document.........................3 2. Introduction...................................................3 3. Problem Description............................................5 3.1. MAC Flush optimization in VPLS resiliency.................5 3.1.1. MAC Flush optimization for regular H-VPLS............5 3.1.2. MAC Flush optimization for native Ethernet access....7 3.2. Black holing issue in PBB-VPLS............................8 4. Solution description...........................................9 4.1. MAC Flush Optimization for VPLS resiliency................9 4.1.1. MAC Flush Parameters TLV format.....................10 4.1.2. Application of MAC Flush TLV in Optimized MAC Flush.11 4.1.3. MAC Flush TLV Processing Rules for regular H-VPLS...12 4.1.4. Optimized MAC Flush Procedures......................12 4.2. LDP MAC Withdraw Extensions for PBB-VPLS.................14 4.2.1. MAC Flush TLV Processing Rules for PBB-VPLS.........15 5. Security Considerations.......................................16 Dutta, et. al Expires December 23, 2011 [Page 2] Internet-Draft Optimized MAC Withdrawal in H-VPLS June 23, 2011 6. IANA Considerations...........................................17 7. Acknowledgments...............................................17 8. References....................................................17 8.1. Normative References.....................................17 8.2. Informative References...................................17 Author's Addresses...............................................18 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 RFC-2119. This document uses the terminology defined in [PBB-VPLS Model], [RFC5036], [RFC4447] and [RFC4762]. Throughout this document VPLS means the emulated bridged LAN service offered to a customer. H-VPLS means the hierarchical connectivity or layout of MTU-s and PE devices offering the VPLS [RFC4762]. The terms spoke node and MTU-s in H-VPLS are used interchangeably. 2. Introduction A method of Virtual Private LAN Service (VPLS), also known as Transparent LAN Service (TLS) is described in [RFC4762]. A VPLS is created using a collection of one or more point-to-point pseudowires (PWs) [RFC4664] configured in a flat, full-mesh topology. The mesh topology provides a LAN segment or broadcast domain that is fully capable of learning and forwarding Ethernet MAC addresses at the PE devices. This VPLS full mesh core configuration can be augmented with additional non-meshed spoke nodes to provide a Hierarchical VPLS (H- VPLS) service [RFC4762]. Throughout this document this configuration is referred to as "regular" H-VPLS. [PBB-VPLS Model] describes how Provider Backbone Bridging (PBB) can be integrated with VPLS to benefit from PBB capabilities while continuing to avoid the use of MSTP in the backbone. The combined solution, referred to as PBB-VPLS, results in better scalability in terms of number of service instances, PWs and customer MACs (CMACs) that need to be handled in the VPLS PEs. Dutta, et. al Expires December 23, 2011 [Page 3] Internet-Draft Optimized MAC Withdrawal in H-VPLS June 23, 2011 A MAC Address Withdrawal mechanism for VPLS is described in [RFC4762] to remove or unlearn MAC addresses for faster convergence on a topology change in resilient H-VPLS topologies. An example of the usage of the MAC Flush mechanism is a dual-homed H-VPLS where an edge device termed as MTU-s is connected to two PE devices via a primary spoke PW and a backup spoke PW, respectively. Such redundancy is designed to protect against the failure of the primary spoke PW or primary PE device. When the MTU-s switches over to the backup PW, it is useful to flush the MAC addresses learned in the corresponding VSI in the peer PE devices participating in the full mesh to avoid black holing of frames to those addresses. Note that a forced switchover to the backup PW can be also performed at MTU-s administratively due to maintenance activities on the primary spoke PW. When the backup PW is made active by the MTU-s it triggers an LDP Address Withdraw Message with a list of MAC addresses to be flushed. The message is forwarded over the LDP session(s) associated with the newly activated PW. In order to minimize the impact on LDP convergence time and scalability when a MAC List TLV contains a large number of MAC addresses, many implementations use an LDP Address Withdraw Message with an empty MAC List. Throughout this document the term MAC Flush Message is used to specify an LDP Address Withdraw Message with an empty MAC List described in [RFC4762] unless specified otherwise. As per the MAC Address Withdrawal processing rules in [RFC4762] a PE device on receiving a MAC flush message removes all MAC addresses associated with the specified VPLS instance (as indicated in the FEC TLV) except for the MAC addresses learned over the newly activated PW. The PE device further triggers a MAC flush message to each remote PE device connected to it in the VPLS full mesh. This method of MAC flushing is modeled after Topology Change Notification (TCN) in Rapid Spanning Tree Protocol (RSTP)[802.1w]. When a bridge switches from a failed link to the backup link, the bridge sends out a TCN message over the newly activated link. The upstream bridge, upon receiving this message, flushes its entire MAC addresses except the ones received over this link and sends the TCN message out of its other ports in that spanning tree instance. The message is further relayed along the spanning tree by the other bridges. When a PE device in the full-mesh of H-VPLS receives a MAC flush message it also flushes MAC addresses which are not affected due to topology change, thus leading to unnecessary flooding and relearning. This document describes the problem and a solution to optimize the MAC flush procedure in [RFC4762] so it flushes the minimal set of MAC addresses that require relearning when the Dutta, et. al Expires December 23, 2011 [Page 4] Internet-Draft Optimized MAC Withdrawal in H-VPLS June 23, 2011 topology changes in H-VPLS. The solution proposed in this document is generic and is applicable when MS-PWs are used in interconnecting PE devices in H-VPLS. [PBB-VPLS Model] describes how PBB can be integrated with VPLS to benefit from PBB capabilities while continuing to avoid the use of MSTP in the backbone. The combined solution referred as PBB-VPLS results in better scalability in terms of number of service instances, PWs and CMACs that need to be handled in the VPLS PEs. This document also describes extensions to LDP MAC Flush procedures described in [RFC4762] required to build desirable capabilities in a PBB-VPLS solution. Section 3 covers the problem space. Section 4 describes the solution and the required TLV extensions. 3. Problem Description 3.1. MAC Flush optimization in VPLS resiliency 3.1.1. MAC Flush optimization for regular H-VPLS Figure 1 describes a dual-homed H-VPLS scenario for a VPLS instance where the problem with the existing MAC flush method in [RFC4762] is explained. Dutta, et. al Expires December 23, 2011 [Page 5] Internet-Draft Optimized MAC Withdrawal in H-VPLS June 23, 2011 PE-1 PE-3 +--------+ +--------+ | | | | | -- | | -- | Customer Site 1 | / \ |------------------| / \ |-> CE-1 /------| \ s/ | | \S / | \ primary spoke PW | -- | /------| -- | \ / +--------+ / +--------+ \ (MTU-s)/ | \ / | +--------+/ | \ / | | | | \ / | | -- | | \ / | | / \ | | H-VPLS Full Mesh Core| | \S / | | / \ | | -- | | / \ | /+--------+\ | / \ | / backup spoke PW | / \ | / \ +--------+ \--------+--------+ CE-2 \ | | | | Customer Site 2 \------| -- | | -- | | / \ |------------------| / \ |-> | \s / | | \S / | | -- | | -- | +--------+ +--------+ PE-2 PE-4 Figure 1: Dual homed MTU-s in two tier hierarchy H-VPLS In Figure 1, the MTU-s is dual-homed to PE-1 and PE-2. Only the primary spoke PW is active at MTU-s, thus PE-1 is acting as the active device to reach the full mesh in the VPLS instance. The MAC addresses of nodes located at access sites (behind CE1 and CE2) are learned at PE-1 over the primary spoke PW. PE-2, PE-3 and PE-4 learn those MAC addresses on their respective mesh PWs terminating to PE-1. Dutta, et. al Expires December 23, 2011 [Page 6] Internet-Draft Optimized MAC Withdrawal in H-VPLS June 23, 2011 When MTU-s switches to the backup spoke PW and activates it, PE-2 becomes the active device to reach the full mesh core. Traffic entering the H-VPLS from CE-1 and CE-2 is diverted by the MTU-s to the spoke PW to PE-2. To avoid traffic black holing the MAC addresses that have been learned in the upstream VPLS full-mesh through PE-1 must be relearned or removed from the MAC FIBs of PE-2, PE-3 and PE- 4. As per the processing rules defined in [RFC4762], on activation of the backup PW from MTU-s, a MAC flush message will be sent by MTU-s to PE-2 that will flush all the MAC addresses learned in the VPLS from all the other PWs except the PWs connected to MTU-s. PE-2 further relays MAC flush messages to all other PE devices in the full mesh. Same processing rule applies at all those PE devices: all the MAC addresses are flushed except the ones learned on the PW to PE2. For example, at PE-3 all of the MAC addresses learned from the PWs connected to PE-1 and PE-4 are flushed and relearned subsequently. Before the relearning happens flooding of unknown destination MAC addresses takes place throughout the network. As the number of PE devices in the full-mesh increases, the number of unaffected MAC addresses flushed in a VPLS instance also increases, thus leading to unnecessary flooding and relearning. With a large number of VPLS instances provisioned in the H-VPLS network topology the amount of unnecessary flooding and relearning increases. An optimization is required that will flush only the MAC addresses learned from the PW connected to PE-1 to minimize the relearning and flooding in the network. Further the forwarding of the MAC Flush by PE-2 delays the overall MAC flush propagation time into the core PEs in the full mesh. So it is desirable to avoid MAC flush forwarding across multiple PEs as far as possible and yet achieve the same desired MAC flushing action. 3.1.2. MAC Flush optimization for native Ethernet access The analysis in section 3.1.1 applies also to the native Ethernet access into a VPLS where one active and one or more backup endpoints into two or more VPLS or H-VPLS PEs are being used. Examples of these are [G.8032v2] access rings or any proprietary multi-chassis LAG emulations. Dutta, et. al Expires December 23, 2011 [Page 7] Internet-Draft Optimized MAC Withdrawal in H-VPLS June 23, 2011 As in the active/standby PWs case from the previous section, upon failure of the active native Ethernet endpoint on PE-1 a MAC Flush optimization is required to ensure that on PE-2, PE-3 and PE-4 only the MAC addresses learned from the PW connected to PE-1 are being flushed. 3.2. Black holing issue in PBB-VPLS In a PBB-VPLS solution a B-component VPLS (B-VPLS) may be used as the infrastructure for one or more I-component instances. B-VPLS control plane (LDP Signaling) replaces the I-component control plane throughout the MPLS core. This raises an additional challenge related to black hole avoidance in the I-component domain as described in this section. Figure 2 describes the case of a CE device (node A) dual-homed to two I-component instances located on two PBB-VPLS PEs (PE1 and PE2). IP/MPLS Core +--------------+ |PE2 | +----+ | |PBB | +-+ | _ |VPLS|---|P| | B/+----+ /+-+\ |PE3 / +----+ / \+----+ +---+/ |PBB |/ +-+ |PBB | +---+ CMAC X--|CE |---|VPLS|---|P|--|VPLS|---|CE |--CMAC Y +---+ A +----+ +-+ +----+ +---+ A |PE1 | B | | +--------------+ Figure 2: PBB Black holing Issue - CE Dual-Homing use case The link between PE1 and CE A is active (marked with A) while the link between CE A and PE2 is in Backup/Blocked status. In the network diagram CMAC X is one of the MAC addresses located behind CE A in the customer domain, CMAC Y is behind CE B and the B-VPLS instances on PE1 and PE2 are associated with backbone MAC (BMAC) B1 and BMAC B2, respectively. As the packets flow from CMAC X to CMAC Y through PE1 with BMAC B1, the remote PEs participating in the I-VPLS (for example, PE3) will Dutta, et. al Expires December 23, 2011 [Page 8] Internet-Draft Optimized MAC Withdrawal in H-VPLS June 23, 2011 learn the CMAC X associated with BMAC B1 on PE1. Under failure of the link between CE A and PE1, and with the activation of link to PE2, the remote PEs (for example, PE3) will black hole the traffic destined to customer MAC X to BMAC B1 until the aging timer expires or a packet flows from X to Y through the PE2 with B2. This may take a long time (default aging timer is 5 minutes) and may affect a large number of flows across multiple I-components. A possible solution to this issue is to use the existing LDP MAC Flush as specified in [RFC4762] to flush in the B-VPLS domain the BMAC associated with the PE where the failure occurred. This will automatically flush the CMAC to BMAC association in the remote PEs. This solution though has the disadvantage of producing a lot of unnecessary MAC flushes in the B-VPLS domain as there was no failure or topology change affecting the Backbone domain. A better solution is required to propagate the I-component events through the backbone infrastructure (B-VPLS) in order to flush only the customer MAC to BMAC entries in the remote PBB-VPLS PEs. As there are no I-VPLS control plane exchanges across the PBB backbone, extensions to the B-VPLS control plane are required to propagate the I-component MAC Flush events across the B-VPLS. 4. Solution description 4.1. MAC Flush Optimization for VPLS resiliency The basic principle of the optimized MAC flush mechanism is explained with reference to Figure 1. PE-1 would initiate a MAC Flush towards the core on detection of the failure of the primary spoke PW between MTU-s and PE-1 (or status change from active to standby). This method is referred to as a PE initiated MAC Flush throughout this document. The MAC Flush message would indicate to receiving PEs to flush all MACs learned over the PW in the context of the VPLS over which the MAC flush message is received. Each PE device in the full mesh that receives the message identifies the VPLS instance and its respective PW that terminates in PE-1 from the FEC TLV received in the message. Thus the PE device flushes only the MAC addresses learned from that PW connected to PE-1 minimizing the required relearning and the flooding throughout the VPLS domain. This section defines a generic MAC Flush Parameters TLV for LDP [RFC5036]. Throughout this document the MAC Flush Parameters TLV is referred to as a MAC Flush TLV. A MAC Flush TLV carries information on the desired action at the PE device receiving the message and is Dutta, et. al Expires December 23, 2011 [Page 9] Internet-Draft Optimized MAC Withdrawal in H-VPLS June 23, 2011 used for optimized MAC flushing in H-VPLS. The MAC Flush TLV is backward compatible and can be used for [RFC4762] style of MAC Flush as explained in section 3.1. 4.1.1. MAC Flush Parameters TLV format The MAC Flush Parameters TLV is described as below: 0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |1|1| MAC Flush Params TLV(TBD) | Length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Flags | Sub-TLV Type | Sub-TLV Length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Sub-TLV Variable Length Value | | " | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ The U and F bits are set to forward if unknown so that potential intermediate VPLS PEs unaware of the new TLV can just propagate it transparently. The MAC Flush Parameters TLV type is to be assigned by IANA. The encoding of the TLV follows the standard LDP TLV encoding in [RFC5036]. The TLV value field contains a one byte Flag field used as described below. Further, the TLV value may carry one or more sub-TLVs. Any sub-TLV definition to the above TLV MUST address the actions in combination with other existing sub-TLVs. The detailed format for the Flags bit vector is described below: 0 1 2 3 4 5 6 7 +-+-+-+-+-+-+-+-+ |C|N| MBZ | (MBZ = MUST Be Zero) +-+-+-+-+-+-+-+-+ 1 Byte Flag field is mandatory. The following flags are defined : C flag, used to indicate the context of the PBB-VPLS component in which MAC flush is required. For PBB-VPLS there are two contexts of MAC flushing - The Backbone VPLS (B-component VPLS) and Customer Dutta, et. al Expires December 23, 2011 [Page 10] Internet-Draft Optimized MAC Withdrawal in H-VPLS June 23, 2011 VPLS (I-component VPLS). C flag MUST be ZERO (C=0) when a MAC Flush for the B-VPLS is required. C flag MUST be set (C=1) when the MAC Flush for I-VPLS is required. In the regular H-VPLS case the C flag MUST be ZERO (C=0) to indicate the flush applies to the current VPLS context. N flag, used to indicate whether a positive (N=0, Flush-all-but- mine) or negative (N=1 Flush-all-from-me) MAC Flush is required. The source (mine/me) is defined either as the PW associated with the LDP session on which the LDP MAC Withdraw was received or with the BMAC(s) listed in the BMAC List Sub-TLV. For the optimized MAC Flush procedure described in this section the flag must be set (N=1). Detailed usage in the context of PBB-VPLS is explained in section 4.2. MBZ flags, the rest of the flags MUST be set to zero on transmission and ignored on reception. 4.1.2. Application of MAC Flush TLV in Optimized MAC Flush For the optimized MAC flush, the MAC Flush TLV MAY be sent as in the existing LDP Address Withdraw Message with an empty MAC List but from the core PE on detection of failure of its local spoke PW. The N bit in the TLV MUST be set to 1. If the optimized MAC Flush procedure is used in a Backbone VPLS or regular VPLS/H-VPLS context the C bit MUST be ZERO (C=0). If it is used in an I-VPLS context the C bit MUST be set (C= 1). See section 4.2 for PBB-VPLS details. Note that if a MAC Flush TLV is not understood by a receiver then it may result in undesired action. For example if a MAC Flush Parameters TLV is received with N=1 and receiver does not understand that TLV then it would result in flushing of all MACs learned in the VSI except the ones learned over the PW. The MAC Flush TLV SHOULD be placed after the existing TLVs in MAC Flush message in [RFC4762]. For backward compatibility of MAC flush initiation procedures as defined in [RFC4762], the PE-1 MAY send a MAC Flush TLV as an OPTIONAL TLV in the MAC Flush Message with N = 0. This would result in same flushing action at the receiving PE devices as desired in [RFC4762]. Dutta, et. al Expires December 23, 2011 [Page 11] Internet-Draft Optimized MAC Withdrawal in H-VPLS June 23, 2011 4.1.3. MAC Flush TLV Processing Rules for regular H-VPLS This section describes the processing rules of a MAC Flush TLV that SHOULD be followed in the context of MAC flush procedures in an H- VPLS. For optimized MAC Flush a multi-homing PE initiates a MAC flush message towards the other related VPLS PEs when it detects a transition (failure or to standby) in an active spoke PW. In such a case the MAC Flush TLV MUST be sent with N = 1. A PE device receiving the MAC Flush TLV SHOULD follow the same processing rules as described in this section. Note that if MS-PW is used in the VPLS then a MAC flush message is processed only at the T-PE nodes since the S-PE(s) traversed by the MS-PW propagate MAC flush messages without any action. In this section, a PE device signifies only T-PE in the MS-PW case unless specified otherwise. When a PE device receives a MAC Flush TLV with N = 1, it SHOULD flush all the MAC addresses learned from the PW in the VPLS in the context on which the MAC Flush message is received. If a MAC Flush TLV is received with N = 0 in the MAC flush message then the receiving PE SHOULD flush the MAC addresses learned from all PWs in the VPLS instance except the ones learned over the PW on which the message is received. If a PE device receives a MAC flush with the MAC Flush TLV option and a valid MAC address list, it SHOULD ignore the option and deal with MAC addresses explicitly as per [RFC4762]. If a PE device that doesn't support MAC Flush TLV receives a MAC flush message with this option, it MUST ignore the option and follow the processing rules as per [RFC4762]. However if the MAC Flush Parameters TLV was sent with N = 1 then it may result in wrong flushing action (Positive MAC Flush). 4.1.4. Optimized MAC Flush Procedures This section explains the optimized MAC flush procedure in the scenario in Figure 1. When the primary spoke PW transition (failure or standby transition) is detected by PE-1, it may send MAC flush messages to PE-2, PE-3 and PE-4 with a MAC Flush TLV and N = 1. Upon receipt of the MAC flush message, PE-2 identifies the VPLS instance that requires the MAC flush from the FEC element in the FEC TLV. On receiving N=1, PE-2 removes all MAC addresses learned from that PW Dutta, et. al Expires December 23, 2011 [Page 12] Internet-Draft Optimized MAC Withdrawal in H-VPLS June 23, 2011 over which the message is received. Same action is followed by PE-3 and PE-4. Figure 3 shows another redundant H-VPLS topology to protect against failure of MTU-s device. Provider RSTP may be used as selection algorithm for active and backup PWs in order to maintain the connectivity between MTU devices and PE devices at the edge. It is assumed that PE devices can detect the failure of PWs in either direction through OAM mechanisms such as VCCV procedures for instance. MTU-1----------------PE-1------------PE-3 | | \ /| | Redundancy | \ / | | Provider RSTP | Full-Mesh . | | | / \ | | | / \| MTU-2================PE-2------------PE-4 Backup PW Figure 3: Redundancy with Provider RSTP MTU-1, MTU-2, PE-1 and PE-2 participate in provider RSTP. By configuration in RSTP it is ensured that the PW between MTU-1 and PE- 1 is active and the PW between MTU-2 and PE-2 is blocked (made backup) at the MTU-2 end. When the active PW failure is detected by RSTP, it activates the PW between MTU-2 and PE-2. When PE-1 detects the failing PW to MTU-1, it may trigger a MAC flush into the full mesh with a MAC Flush TLV that carries N=1. Other PE devices in the full mesh that receive the MAC flush message identify their respective PWs terminating on PE-1 and flush all the MAC addresses learned from it. By default, MTU-2 should still trigger MAC flush as currently defined in [RFC4762] after the backup PW is made active by RSTP. Mechanisms to prevent two copies of MAC withdraws to be sent in such scenarios is out of scope of this document. [RFC4762] describes multi-domain VPLS services where fully meshed VPLS networks (domains) are connected together by a single spoke PW per VPLS service between the VPLS "border" PE devices. To provide redundancy against failure of the inter-domain spoke, a full mesh of Dutta, et. al Expires December 23, 2011 [Page 13] Internet-Draft Optimized MAC Withdrawal in H-VPLS June 23, 2011 inter-domain spokes can be setup between border PE devices and provider RSTP may be used for selection of the active inter-domain spoke. In case of an inter-domain spoke PW failure, PE initiated MAC withdrawal may be used for optimized MAC flushing within individual domains. Further, the procedures are applicable with any native Ethernet access topologies multi-homed to two or more VPLS PEs. The text in section 4.1 applies for the native Ethernet case where active/standby PWs are replaced with the active/standby Ethernet endpoints. An optimized MAC Flush message can be generated by the VPLS-PE that detects the failure in the primary Ethernet access. 4.2. LDP MAC Withdraw Extensions for PBB-VPLS The use of Address Withdraw messages with MAC List TLV is proposed in [RFC4762] as a way to expedite removal of MAC addresses as the result of a topology change (e.g. failure of a primary link of a VPLS PE and implicitly the activation of an alternate link in a dual-homing use case). These existing procedures apply individually to B-VPLS and I- component domains. When it comes to reflecting topology changes in access networks connected to I-component across the B-VPLS domain certain additions should be considered as described below. MAC Switching in PBB is based on the mapping of Customer MACs (CMACs) to Backbone MAC(s) (BMACs). A topology change in the access (I- domain) should just invoke the flushing of CMAC entries in PBB PEs' FIB(s) associated with the I-component(s) impacted by the failure. There is a need to indicate the PBB PE (BMAC source) that originated the MAC Flush message to selectively flush only the MACs that are affected. These goals can be achieved by adding a new MAC Flush Parameters TLV in the LDP Address Withdraw message to indicate the particular domain(s) requiring MAC flush. On the other end, the receiving PEs may use the information from the new TLV to flush only the related FIB entry/entries in the I-component instance(s). The following sub-TLVs MUST be included in the MAC Flush Parameters TLV if the C-flag is set to 1: Dutta, et. al Expires December 23, 2011 [Page 14] Internet-Draft Optimized MAC Withdrawal in H-VPLS June 23, 2011 - PBB BMAC List sub-TLV: Type: 0x01 Length: value length in octets. At least one BMAC address must be present in the list. Value: one or a list of 48 bits BMAC addresses. These are the source BMAC addresses associated with the B-VPLS instance that originated the MAC Withdraw message. It will be used to identify the CMAC(s) mapped to the BMAC(s) listed in the sub-TLV. - PBB ISID List sub-TLV: Type: 0x02, Length: value length in octets. Zero indicates an empty ISID list. An empty ISID list means that the flush applies to all the ISIDs mapped to the B-VPLS indicated by the FEC TLV. Value: one or a list of 24 bits ISIDs that represent the I-component FIB(s) where the MAC Flush needs to take place. 4.2.1. MAC Flush TLV Processing Rules for PBB-VPLS The following steps describe the details of the processing for the related LDP Address Withdraw message: . The LDP MAC Withdraw Message, including the MAC Flush Parameters TLV is initiated by the PBB PE(s) experiencing a Topology Change event in one or multiple customer I-component(s). o The flags are set accordingly to indicate the type of MAC Flush required for this event: N=0 (Flush-all-but-mine) or N = 1 (Flush-all-from-me), C=1 (Flush only CMAC FIBs). o The PBB Sub-TLVs (BMAC and ISID Lists) are included according to the context of topology change. . On reception of the LDP Address Withdrawal message, the B-VPLS instances corresponding to the FEC TLV in the message must interpret the content of MAC Flush Parameters TLV. If the C-bit is set to 1 then Backbone Core Bridges (BCB) in the PBB-VPLS SHOULD NOT flush their BMAC FIBs. The B-VPLS control plane SHOULD propagate the MAC Flush following the split-horizon grouping and the established B-VPLS topology. Dutta, et. al Expires December 23, 2011 [Page 15] Internet-Draft Optimized MAC Withdrawal in H-VPLS June 23, 2011 . The usage and processing rules of MAC Flush Parameters TLV in the context of Backbone Edge Bridges (BEB) is as follows: o The PBB ISID List is used to determine the particular ISID FIBs (I-VPLS) that need to be flushed. If the ISID List is empty then all the ISID FIBs associated with the receiving B-VPLS SHOULD be flushed. o The PBB BMAC List is used to identify from the ISID FIBs in the previous step whether to selectively flush BMAC to CMAC associations depending on the N flag specified below. . Next, depending on the N flag value the following actions apply: o N=0, all the CMACs in the selected ISID FIBs SHOULD be flushed with the exception of the identified CMAC list from the BMAC List mentioned in the message. ("Flush all but the CMACs associated with the BMAC(s) in the BMAC List Sub-TLV from the FIBs associated with the ISID list"). o N=1, the identified CMAC list SHOULD be flushed ("Flush all the CMACs associated with the BMAC(s) in the BMAC List Sub- TLV from the FIBs associated with the ISID list"). 4.2.3 Applicability of MAC Flush Parameters TLV If a MAC Flush Parameters TLV is received by a BEB in a PBB- VPLS that does not understand the TLV then it may result in undesirable MAC flushing action. It is RECOMMENDED that all PE devices participating in PBB-VPLS support MAC Flush Parameters TLV. The MAC Flush Parameters TLV is also applicable to regular VPLS contexts. To achieve negative MAC Flush (flush-all-from-me) in a regular VPLS context, the MAC Flush Parameters TLV SHOULD be encoded with C=0 and N = 1 without the inclusion of any Sub-TLVs. Negative MAC flush is highly desirable in scenarios when VPLS access redundancy is provided by Ethernet Ring Protection as specified in [G.8032v2] specification etc. 5. Security Considerations Control plane aspects: - LDP security (authentication) methods as described in [RFC5036] is applicable here. Further this document implements security considerations as in [RFC4447] and [RFC4762]. Dutta, et. al Expires December 23, 2011 [Page 16] Internet-Draft Optimized MAC Withdrawal in H-VPLS June 23, 2011 Data plane aspects: - This specification does not have any impact on the VPLS forwarding plane. 6. IANA Considerations The Type field in MAC Flush Parameters TLV is defined as 0x406 and is subject to IANA approval. 7. Acknowledgments The authors would like to thank the following people who have provided valuable comments and feedback on the topics discussed in this document: Marc Lasserre, Ian Cowburn, Dimitri Papadimitriou, Jorge Rabadan, Prashanth Ishwar, Vipin Jain, John Rigby, Ali Sajassi, Wim Henderickx, Jorge Rabadan and Maarten Vissers. 8. References 8.1. Normative References [RFC4762] Lasserre, M. and Kompella, V. (Editors), "Virtual Private LAN Service (VPLS) Using Label Distribution Protocol (LDP) Signaling", RFC 4762, January 2007. [RFC5036] Andersson, L., et al. "LDP Specification", RFC5036, October 2007. [RFC4447] Martini. and et al., "Pseudowire Setup and Maintenance Using Label Distribution Protocol (LDP)", RFC 4447, April 2006. 8.2. Informative References [PBB-VPLS Model] F. Balus, et Al. "Extensions to VPLS PE model for Provider Backbone Bridging", draft-ietf-l2vpn-pbb-vpls-pe- model-00.txt, May 2009 (work in progress) [RFC4664] Andersson, L., et al. "Framework for Layer 2 Virtual Private Networks (L2VPNs)", RFC 4664, September 2006. Dutta, et. al Expires December 23, 2011 [Page 17] Internet-Draft Optimized MAC Withdrawal in H-VPLS June 23, 2011 [802.1w] "IEEE Standard for Local and metropolitan area networks. Common specifications Part 3: Media Access Control (MAC) Bridges. Amendment 2: Rapid Reconfiguration", IEEE Std 802.1w-2001. [G.8032v2] ITU-T G.8032v2 specification Author's Addresses Pranjal Kumar Dutta Alcatel-Lucent 701 E Middlefield Road, Mountain View, CA 94043 USA Email: pranjal.dutta@alcatel-lucent.com Florin Balus Alcatel-Lucent 701 E. Middlefield Road Mountain View, CA, USA 94043 Email: florin.balus@alcatel-lucent.com Geraldine Calvignac France Telecom 2, avenue Pierre-Marzin 22307 Lannion Cedex France Email: geraldine.calvignac@orange-ftgroup.com Olen Stokes Extreme Networks PO Box 14129 RTP, NC 27709 USA Email: ostokes@extremenetworks.com Dutta, et. al Expires December 23, 2011 [Page 18]