Network Working Group Raymond Key, Huawei Internet Draft Simon Delord, Alcatel-Lucent Category: Informational Frederic Jounay, Orange France Telecom Expires: April 2012 Lucy Yong, Huawei Lizhong Jin, ZTE Yuji Kamite, NTT Communications Wim Henderickx, Alcatel-Lucent October 16, 2011 A Framework for E-Tree Service over MPLS Network draft-key-l2vpn-etree-frwk-06 Status of this Memo This Internet-Draft is submitted to IETF 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 April 16, 2012. Abstract This document proposes a solution framework for supporting Metro Ethernet Forum (MEF) Ethernet Tree (E-Tree) services over a Multiprotocol Label Switching (MPLS) network. The objective is to provide a simple and effective approach to emulate E-Tree services in addition to Ethernet LAN (E-LAN) services on an existing MPLS network. Key, et al. Expires April 2012 [Page 1]
Internet Draft Framework E-Tree over MPLS October 2011 Table of Contents 1. Introduction....................................................3 1.1. Objective and Scope...........................................3 1.2. Traditional Ethernet Network..................................3 1.3. MEF Multipoint Ethernet Services..............................3 1.3.1. Similarity between E-LAN and E-Tree.........................4 1.3.2. Difference between E-LAN and E-Tree.........................4 1.4. IETF Multipoint L2VPN Services................................5 1.4.1. Virtual Private LAN Service (VPLS)..........................5 1.4.2. Virtual Private Multicast Service (VPMS)....................5 1.5. Terminology...................................................6 2. Reference Model.................................................6 3. Use Cases.......................................................8 4. Challenges......................................................9 4.1. Generic E-Tree Service Definition.............................9 4.1.1. Mandatory Leaf-to-Leaf Communication Restriction............9 4.2. Use Case Desirable Requirements..............................10 4.2.1. Ethernet Broadcast/Multicast Optimisation..................10 4.2.2. IP Multicast Optimisation..................................11 4.2.3. MAC-based Forwarding Unnecessary...........................11 4.2.4. MAC-based Forwarding Security Concern......................12 5. A Solution Framework for MAC-based Forwarding E-Tree...........12 5.1. MAC-based Forwarding Any-to-Any Ethernet VPN.................12 5.2. Leaf-to-Leaf Communication Restriction.......................13 5.3. Optional Enhancement - Point-to-Multipoint PW................13 5.4. Optional Enhancement - IP Multicast in VPLS.......... .......14 6. Non-MAC-based Forwarding E-Tree................................14 6.1. Single Root, Broadcast Only - VPMS...........................14 6.2. Multiple Roots, Broadcast and Unicast........................14 7. Security Consideration.........................................14 8. IANA Considerations............................................15 9. Acknowledgements...............................................15 10. References....................................................15 10.1. Normative References........................................15 10.2. Informative References......................................15 Appendix A. Some Possible Ways for Leaf-to-Leaf Communication Restriction...........................................17 Authors' Addresses................................................27 Intellectual Property and Copyright Statements....................28 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]. Key, et al. Expires April 2012 [Page 2]
Internet Draft Framework E-Tree over MPLS October 2011 1. Introduction 1.1. Objective and Scope This document proposes a solution framework for supporting Metro Ethernet Forum (MEF) Ethernet Tree (E-Tree) services over a MPLS network. The objective is to provide a simple and effective approach to emulate E-Tree services in addition to Ethernet LAN (E-LAN) services on an existing MPLS network. This solution framework makes use of existing IETF specified mechanisms unless there are technical reasons why the existing mechanisms are insufficient or unnecessary. This document does not intend to provide a full specification of the solution, but rather to identify the functional components of the overall solution, and for each component, whether it is REQUIRED or OPTIONAL, whether existing mechanism is sufficient, or whether relevant mechanism is already under development. In this document, "current standard" refers to [RFC4385], [RFC4447], [RFC4448], [RFC4761] and [RFC4762]. 1.2. Traditional Ethernet Network In this document, traditional Ethernet network refers to the Ethernet bridge/switch network, not the Ethernet repeater/hub network. Data frame is Ethernet frame. Data forwarding is MAC-based forwarding, which includes MAC address learning and aging. It is important to note that in traditional Ethernet network unicast unknown, multicast and broadcast frames are forwarded in exactly the same way to every port except the ingress port. An Ethernet host receiving a frame checks the destination address in the frame to decide whether it is the intended destination. 1.3. MEF Multipoint Ethernet Services MEF defines two multipoint Ethernet Service types: - E-LAN (Ethernet LAN), multipoint-to-multipoint service - E-Tree (Ethernet Tree), rooted-multipoint service According to MEF's technical specification, a generic E-LAN/E-Tree service is always bidirectional in the sense that ingress frames can originate at any endpoint in the service. However, some application scenarios of E-Tree may have unidirectional traffic only. Section 3 will discuss about different use cases. Key, et al. Expires April 2012 [Page 3]
Internet Draft Framework E-Tree over MPLS October 2011 For full specification, please refer to MEF's "Ethernet Services Definitions - Phase 2" [MEF6.1] and "Ethernet Services Attributes Phase 2" [MEF10.2]. 1.3.1. Similarity between E-LAN and E-Tree Data frame is Ethernet frame. Data forwarding can be MAC-based forwarding or something else, to be specified by service provider in the particular service definition. Extract from [MEF6.1] Table 7 and Table 9: +---------------+---------------------------------------------------+ | EVC Service | E-LAN/E-Tree Service Type Requirement | | Attribute | | +---------------+---------------------------------------------------+ | Unicast | Deliver Unconditionally or Deliver Conditionally. | | Service Frame | If Delivered Conditionally, MUST specify the | | Delivery | delivery criteria. | +---------------+---------------------------------------------------+ | Multicast | Deliver Unconditionally or Deliver Conditionally. | | Service Frame | If Delivered Conditionally, MUST specify the | | Delivery | delivery criteria. | +---------------+---------------------------------------------------+ | Broadcast | Deliver Unconditionally or Deliver Conditionally. | | Service Frame | If Delivered Conditionally, MUST specify the | | Delivery | delivery criteria. | +---------------+---------------------------------------------------+ It is important to note that it is not a must for a MEF multipoint Ethernet service (E-LAN or E-Tree) to use MAC-based forwarding. This document presents a solution framework for MAC-based forwarding E-Tree in section 5, and also discusses non-MAC-based forwarding E-Tree in section 6. 1.3.2. Difference between E-LAN and E-Tree Within the context of a multipoint Ethernet service, each endpoint is designated as either a Root or a Leaf. A Root can communicate with all other endpoints in the same multipoint Ethernet service, however a Leaf can only communicate with Roots but not Leafs. The only difference between E-LAN and E-Tree is: - E-LAN has Root endpoints only, which implies there is no communication restriction between endpoints - E-Tree has both Root and Leaf endpoints, which implies there is a need to enforce communication restriction between Leaf endpoints Extract from [MEF10.2] Section 6.3: The UNI Type MUST have the value either "Root" or "Leaf." If the type of EVC is Point-to-Point or Multipoint-to-Multipoint, then the UNI Type MUST equal "Root." Key, et al. Expires April 2012 [Page 4]
Internet Draft Framework E-Tree over MPLS October 2011 Extract from [MEF10.2] Section 6.1.2.2: An ingress Service Frame mapped to the EVC at a Leaf UNI MUST NOT result in an egress Service Frame at another Leaf UNI but MAY result in an egress Service Frame at some or all of the Root UNIs. It is important to note that one E-Tree service may have single or multiple Root UNIs. Extract from [MEF6.1] Section 6.3: In its simplest form, an E-Tree Service type can provide a single Root for multiple Leaf UNIs. Each Leaf UNI can exchange data with only the Root UNI. ... In more sophisticated forms, an E-Tree Service type may support two or more Root UNIs. In this scenario, each Leaf UNI can exchange data only with the Root UNIs. As well, the Roots can communicate with each other. In such a service, redundant access to the Root can also be provided, effectively allowing for enhanced service reliability and flexibility. 1.4. IETF Multipoint L2VPN Services 1.4.1. Virtual Private LAN Service (VPLS) VPLS is a L2VPN service that provides multipoint-to-multipoint connectivity for Ethernet across an IP or MPLS-enabled IP Packet Switched Network. VPLS emulates the Ethernet VLAN functionality of traditional Ethernet network. VPLS is a current IETF standard, please refer to [RFC4761] [RFC4762]. Data frame is Ethernet frame. Data forwarding is MAC-based forwarding, which includes MAC address learning and aging. It is important to note that the current standard VPLS treats Ethernet multicast frame in exactly the same way as Ethernet broadcast frame and does not restrict transmission of Ethernet multicast frame to a smaller set of receivers. An Ethernet host receiving a frame checks the destination address in the frame to determine whether it is the intended destination. VPLS can be used to emulate E-LAN service over MPLS network provided that the E-LAN service uses MAC-based forwarding as service frame delivery attribute. Considerable number of service providers have adopted this approach to provide E-LAN services to customers. 1.4.2. Virtual Private Multicast Service (VPMS) VPMS is a L2VPN service that provides point-to-multipoint connectivity across a variety of link layers, including Frame Relay, ATM, Ethernet, PPP, etc., across an IP or MPLS-enabled IP Packet Switched Network. Key, et al. Expires April 2012 [Page 5]
Internet Draft Framework E-Tree over MPLS October 2011 In the Ethernet use case, VPMS provides single coverage of receiver membership, i.e. there is no distinct differentiation for multiple multicast groups. Destination address in Ethernet frame is not used in data forwarding. VPMS MUST support unidirectional point-to-multipoint traffic from a sender to multiple receivers and MAY support reverse traffic in a point-to-point manner. VPMS is currently under development. Please refer to [Draft VPMS Frmwk]. 1.5. Terminology E-Tree An Ethernet VPN in which each Root AC can communicate with every other AC, whereas Leaf ACs can only communicate with Root ACs. Each AC on an E-Tree construct is designated as either a Root AC or a Leaf AC. There can be multiple Root ACs and Leaf ACs per E-Tree construct. Root AC An ingress frame at a Root AC can be delivered to one or more of any of the other ACs in the E-Tree. Please note that this AC is bidirectional. Leaf AC Ingress frame at a Leaf AC can only be delivered to one or more Root ACs in the E-Tree. Ingress frame at a Leaf AC MUST NOT be delivered to any Leaf ACs in the E-Tree. Please note that this AC is bidirectional. 2. Reference Model Figure 1 below describes a generic reference model where PE1, PE2 and PE3 need to establish an E-Tree construct between different Ethernet endpoints. Each PE has 2 Root ACs and 2 Leaf ACs connected to a VSI. These VSIs are then linked together via Ethernet PWs. In most use cases, an E-Tree construct has only a few Root ACs but many Leaf ACs. There may be only Root ACs or only Leaf ACs on a PE. Key, et al. Expires April 2012 [Page 6]
Internet Draft Framework E-Tree over MPLS October 2011 <------------E-Tree------------> +---------+ +---------+ | PE1 | | PE2 | +----+ | +---+ | | +---+ | +----+ |CE01+----AC1----+--+ | | | | +--+----AC5----+CE05| +----+ (Root AC) | | V | | | | V | | (Root AC) +----+ +----+ | | | | | | | | +----+ |CE02+----AC2----+--+ | | Ethernet | | +--+----AC6----+CE06| +----+ (Root AC) | | S +--+-----PW-----+--+ S | | (Root AC) +----+ +----+ | | | | | | | | +----+ |CE03+----AC3----+--+ | | | | +--+----AC7----+CE07| +----+ (Leaf AC) | | I | | | | I | | (Leaf AC) +----+ +----+ | | | | | | | | +----+ |CE04+----AC4----+--+ | | | | +--+----AC8----+CE08| +----+ (Leaf AC) | +-+-+ | | +-+-+ | (Leaf AC) +----+ | | | | | | +----+----+ +----+----+ | | |Ethernet |Ethernet |PW |PW | | | +----+----+ | | | | | | +-+-+ | +----+ | | | +--+----AC9----+CE09| | | | V | | (Root AC) +----+ | | | | | +----+ | | | +--+----AC10---+CE10| +-----------------+--+ S | | (Root AC) +----+ | | | | +----+ | | +--+----AC11---+CE11| | | I | | (Leaf AC) +----+ | | | | +----+ | | +--+----AC12---+CE12| | +---+ | (Leaf AC) +----+ | PE3 | +---------+ <------------E-Tree------------> Figure 1: E-Tree Reference Model With an E-Tree construct: - A Root AC can receive from and transmit to any other ACs. - A Leaf AC can receive from and transmit to any Root ACs. - A Leaf AC cannot receive from and transmit to any other Leaf ACs. This applies to all traffic, including Unicast Known, Unicast Unknown, Broadcast and Multicast. When an Ethernet Frame is received on PE1 via AC1, the frame can be transmitted to any other local ACs on PE1 and via Ethernet PWs to any remote ACs on PE2 and PE3. Key, et al. Expires April 2012 [Page 7]
Internet Draft Framework E-Tree over MPLS October 2011 However when an Ethernet frame is received on PE1 via AC3, the frame can be transmitted to any other local Root ACs on PE1 and via Ethernet PWs to any remote Root ACs on PE2 and PE3, but the frame cannot be transmitted to any local Leaf ACs on PE1 nor any remote Leaf ACs on PE2 and PE3. 3. Use Cases Table 1 below presents some major use cases. +---------------------------+--------------+------------+ | Use Case | Root | Leaf | +---+---------------------------+--------------+------------+ | 1 | Hub & Spoke VPN | Hub Site | Spoke Site | +---+---------------------------+--------------+------------+ | 2 | Wholesale Access | Customer's | Customer's | | | | Interconnect | Subscriber | +---+---------------------------+--------------+------------+ | 3 | Mobile Backhaul | RAN NC | RAN BS | +---+---------------------------+--------------+------------+ | 4 | IEEE 1588 PTPv2 | PTP Server | PTP Client | | | Clock Synchronisation | | | +---+---------------------------+--------------+------------+ | 5 | Internet Access | BNG Router | Subscriber | +---+---------------------------+--------------+------------+ | 6 | Broadcast Video | Video Source | Subscriber | | | (unidirectional only) | | | +---+---------------------------+--------------+------------+ | 7 | Broadcast/Multicast Video | Video Source | Subscriber | | | plus Control Channel | | | +---+---------------------------+--------------+------------+ | 8 | Device Management | Management | Managed | | | | System | Device | +---+---------------------------+--------------+------------+ Table 1: E-Tree Use cases Common to all use cases, direct Layer 2 Leaf-to-Leaf communication is not required. For Mobile backhaul, this may not be valid for LTE X2 interfaces in the future. If direct Layer 2 Leaf-to-Leaf communication is not allowed due to security concern, then E-Tree should be used to prohibit communication between Leaf endpoints, otherwise E-LAN is also a feasible option. Also common to the use cases mentioned above, there may be single or multiple Root endpoints in one E-Tree service. The need for multiple Root endpoints is usually driven by redundancy requirement. Whether a particular E-Tree service needs to support single or multiple Root endpoints depends on the target application. Key, et al. Expires April 2012 [Page 8]
Internet Draft Framework E-Tree over MPLS October 2011 A generic E-Tree service supports all the following traffic flows: - Ethernet Unicast from Root to Leaf - Ethernet Unicast from Leaf to Root - Ethernet Unicast from Root to Root - Ethernet Broadcast/Multicast from Root to Roots & Leafs - Ethernet Broadcast/Multicast from Leaf to Roots A particular E-Tree service may need to support all the above or only a subset depending on the target application. Among the use cases mentioned above, broadcast video draws most attention. Actually, broadcast video is a representing example for content delivery in general, such as news feed, financial data feed, etc. 4. Challenges 4.1. Generic E-Tree Service Definition This section highlights why the current standard VPLS is insufficient for emulating E-Tree service over MPLS network. 4.1.1. Mandatory Leaf-to-Leaf Communication Restriction Current standard VPLS treats all ACs equal (i.e. not classified into Root or Leaf) and provides any-to-any connectivity among all ACs. The current standard VPLS does not include any mechanism of communication restriction between specific ACs, therefore is insufficient for emulating generic E-Tree service over MPLS network. A problem occurs when there are two or more PEs with both Root AC and Leaf AC. Let's look at the scenario illustrated in Figure 2 below. VPLS is used to emulate an E-Tree service over a MPLS network. Note: Figure 2 is a hypothetical case solely for explaining the problem, and not meant to represent a typical E-Tree service. <------------E-Tree------------> +---------+ +---------+ | PE1 | | PE2 | +---+ | +---+ | | +---+ | +---+ |CE1+-----AC1----+--+ | | | | +--+----AC3-----+CE3| +---+ (Root AC) | | V | | Ethernet | | V | | (Root AC) +---+ | | S +--+-----PW-----+--+ S | | +---+ | | I | | | | I | | +---+ |CE2+-----AC2----+--+ | | | | +--+----AC4-----+CE4| +---+ (Leaf AC) | +---+ | | +---+ | (Leaf AC) +---+ +---------+ +---------+ Figure 2: Problem Scenario for Leaf-to-Leaf Communication Restriction Key, et al. Expires April 2012 [Page 9]
Internet Draft Framework E-Tree over MPLS October 2011 When PE2 receives a frame from PE1 via the Ethernet PW, - PE2 does not know which AC on PE1 is the ingress AC - PE2 does not know whether the ingress AC is a Leaf AC or not - PE2 does not have sufficient information to enforce the Leaf-to-Leaf communication restriction Examples: - CE2 sends a Broadcast/Multicast frame to PE1 via AC2 - CE2 sends a Unicast frame to PE1 via AC2, destination address in Ethernet header equal to CE4's MAC address In order to fulfil the generic E-Tree service definition, extension to the current VPLS standard will be required. Extension to related PWE3 standard may also be required, depending on solution approach. Such extensions should have minimal impact on the emulated E-LAN services already in operation. There are some possible ways to get around this problem that do not require extension to the current VPLS standard but they all come with significant design complexity or deployment constraints. Appendix A highlights the major ones and the related concerns. 4.2. Use Case Desirable Requirements There are quite a variety of use cases for E-Tree. For some use cases, the generic MEF E-Tree service definition is good enough. For some other use cases, there are desirable requirements beyond that. The challenges discussed in this section are not related to the generic MEF E-Tree service definition but the desirable requirements of specific use cases. They may be critical to the success in some E-Tree services while totally irrelevant in some others. 4.2.1. Ethernet Broadcast/Multicast Optimisation According to MAC-based forwarding, an Ethernet broadcast/multicast/ unicast unknown frame is forwarded to all ACs other than the ingress AC, which implies point-to-multipoint traffic from the ingress PE to all other PEs in the VPLS instance. The current standard VPLS uses only point-to-point PW between PEs. When the Ethernet destination address is broadcast, multicast or unicast unknown, the ingress PE replicates the frame on every PW towards remote PE belonging to the same VPLS instance. Depending on the mapping between the logical topology of the E-Tree service and the physical topology of the network, multiple PWs may transverse same physical link, result in multiple copies of the same payload Ethernet frame on the physical link. Such approach is inefficient in terms of bandwidth usage. Key, et al. Expires April 2012 [Page 10]
Internet Draft Framework E-Tree over MPLS October 2011 For some use cases, for example broadcast/multicast video, due to nature of the application, there is significant volume of point-to- multipoint traffic. Bandwidth optimisation for such traffic within the network becomes a concern from the service provider perspective. [RFC5501] provides an in-depth discussion on broadcast/multicast related requirements for VPLS, see issue B (Replication of PWs on shared physical path) in section 3.2. 4.2.2. IP Multicast Optimisation The current standard VPLS is a L2VPN service agnostic to customer's Layer 3 traffic, hence does not maintain any information about IP multicast group membership. Although a Layer 3 IP multicast packet is encapsulated in a Layer 2 Ethernet multicast frame, the current standard VPLS treats Ethernet multicast frame in exactly the same way as Ethernet broadcast frame. Therefore, such payload IP multicast packet will be forwarded to every other AC of the same VPLS instance. A payload IP multicast packet will be forwarded to all ACs, including those with no member of the specific IP multicast group attached. Unnecessary traffic consumes bandwidth on access link and may become a concern from the customer perspective. In some cases, it may also be a security concern as the multicast frame may be forwarded to an endpoint other than the intended destinations. A payload IP multicast packet will be forwarded to a remote PE with no member of the specific IP multicast group attached. Unnecessary traffic consumes bandwidth in the network and may become a concern from the service provider perspective. For some use cases, for example multicast video, due to nature of the application, there is significant volume of IP multicast traffic and different IP multicast groups are required in one E-Tree service. The above may become a real concern from both the customer and service provider perspectives. [RFC5501] provides an in-depth discussion on broadcast/multicast related requirements for VPLS, see both issue A (Replication to non- member site) and issue B (Replication of PWs on shared physical path) in section 3.2. 4.2.3. MAC-based Forwarding Unnecessary For some use cases, for example broadcast video, due to nature of the application, there is only broadcast unidirectional traffic from Root to all other endpoints. It is unnecessary to use destination address for data forwarding. Deliver unconditionally for ingress frame at Root endpoint may be a simpler approach than MAC-based forwarding. Key, et al. Expires April 2012 [Page 11]
Internet Draft Framework E-Tree over MPLS October 2011 4.2.4. MAC-based Forwarding Security Concern MAC-based forwarding will make an unicast frame from a Root destined for a specific Leaf being forwarded to other endpoints in addition to the intended destination when the frame is classified as unicast unknown, may be due to MAC address aged out or MAC address table overflow. MAC address spoofing may cause an unicast frame from a Root destined for a specific Leaf being forwarded to an endpoint different from the intended destination. If such unicast frame carries sensitive information strictly for the intended destination only, then the MAC-based forwarding may cause a security concern from the customer perspective. For some use cases where mutually un-trusted subscribers are connected to leaf endpoints in the same E-Tree service, such as Internet access and wholesale access, this is a valid concern. There are some possible mitigations: - For every Leaf endpoint of the particular E-Tree service, deploy a service provider controlled router between the Leaf endpoint and the customer network - Customer to deploy encryption for sensitive information, for example IPsec, SSL, SSH, HTTPS Whether the MAC-based forwarding really becomes a security concern depends on the particular application and the deployment scenario. This is unlikely to be a critical concern in most cases. 5. A Solution Framework for MAC-based Forwarding E-Tree As mentioned in section 1.3.1. E-Tree can use MAC-based forwarding or something else for data forwarding. This section presents a solution framework for MAC-based forwarding E-Tree. Section 6 will discuss other variants. This is a VPLS-based solution. Functional components of the solution are identified and discussed in the subsections. 5.1. MAC-based Forwarding Any-to-Any Ethernet VPN This is a REQUIRED component. This component is the current standard VPLS and PWE3 as specified in [RFC4385] [RFC4447] [RFC4448] [RFC4761] [RFC4762], which provides any-to-any connectivity among all ACs in one VPLS instance. This is the base component. All other REQUIRED/OPTIONAL components are to be added on top of this component. Key, et al. Expires April 2012 [Page 12]
Internet Draft Framework E-Tree over MPLS October 2011 5.2. Leaf-to-Leaf Communication Restriction This is in response to the challenge in section 4.1.1. Mandatory Leaf-to-Leaf Communication Restriction. This is a REQUIRED component. This component is a minimal extension to the current VPLS and PWE3 standards, with the objective to provide a simple and effective way to support generic E-Tree services in addition to E-LAN services using VPLS on a MPLS network. [Draft VPLS ETree Req] is a work in progress requirement draft. Different solutions have been proposed: - Control Word L-bit solution, [Draft CW L-bit] [Draft VPLS ETree] - Dual VLAN solution, [Draft VPLS PE ETree] - Two PW solution, [Draft VPLS ETree 2PW] - Two VE ID solution [Draft VPLS ETree 2VEID] 5.3. Optional Enhancement - Point-to-Multipoint PW This is in response to the challenge in section 4.2.1. Ethernet Broadcast/Multicast Optimisation. This is an OPTIONAL component, applicable only when there is significant volume of Ethernet broadcast/multicast traffic. Point-to-Multipoint pseudowire (P2MP PW) is a PW attached to a source used to distribute Layer 1 or Layer 2 format traffic to a set of receivers. P2MP PW is unidirectional but optionally bidirectional. By using P2MP PW, the ingress PE is not responsible for replicating the payload frame on each P2P PW towards egress PE, instead the network elements along the physical path participate in replication. The replication is done by the underlying point-to-multipoint label switched path (P2MP LSP). Extension to current VPLS standard will be required to specify how P2MP PW and P2P PW should be used and how MAC learning works on P2MP PW. Please refer to [Draft LDP-VPLS Bcast]. P2MP PW is currently under development. Please refer to [Draft P2MP PW Req] [Draft P2MP PW Sig]. It is important to note that this component will align with the recommendation in [RFC4665], "With the exception of IPLS, an L2VPN service SHOULD be agnostic to customer's Layer 3 traffic (e.g., IP, IPX, Appletalk) encapsulated within Layer 2 frames." Key, et al. Expires April 2012 [Page 13]
Internet Draft Framework E-Tree over MPLS October 2011 5.4. Optional Enhancement - IP Multicast in VPLS This is in response to the challenge in section 4.2.2. IP Multicast Optimisation. This is an OPTIONAL component, applicable only when there is significant volume of IP multicast traffic and different IP multicast groups are required in one E-Tree service. Multicast in VPLS is currently under development, with the objective to provide efficient ways to support IP multicast services over VPLS. It covers IP multicast group membership control and also bandwidth optimisation. Please refer to [Draft Mcast VPLS]. It is important to note that this component will make use of Layer 3 IP multicast information in payload frames to improve transport efficiency, hence will not align with the recommendation in [RFC4665] that an L2VPN service SHOULD be agnostic to customer's Layer 3 traffic. 6. Non-MAC-based Forwarding E-Tree This section presents some variants of E-Tree services which do not use MAC-based forwarding as the service frame delivery attribute. 6.1. Single Root, Broadcast Only - VPMS This is in response to the challenge in section 4.2.3. MAC-based Forwarding Unnecessary. VPMS provides single coverage of receiver membership. Destination address in Ethernet frame is not used in data forwarding. For E-Tree service of single Root and only unidirectional broadcast traffic from the Root, for example certain broadcast video or similar content delivery applications, VPMS will be a much more simple and effective solution than VPLS. VPMS is currently under development. Please refer to [Draft VPMS Frmwk]. 6.2. Multiple Roots, Broadcast and Unicast This is in response to the challenge in section 4.2.4. MAC-based Forwarding Security Concern. This will be added in later version of this document. 7. Security Considerations This will be added in later version of this document. Key, et al. Expires April 2012 [Page 14]
Internet Draft Framework E-Tree over MPLS October 2011 8. IANA Considerations This will be added in later version of this document. 9. Acknowledgements This will be added in later version of this document. 10. References 10.1. Normative References [MEF6.1] Metro Ethernet Forum, Ethernet Services Definitions - Phase 2, April 2008 [MEF10.2] Metro Ethernet Forum, Ethernet Services Attributes Phase 2, October 2009 [RFC2119] Bradner, S., Key words for use in RFCs to Indicate Requirement Levels, BCP 14, RFC 2119, March 1997 [RFC4385] Bryant,S., Swallow, G., and Al, Pseudowire Emulation Edge-to-Edge (PWE3) Control Word for Use over an MPLS PSN, February 2006. [RFC4447] Martini, L., and al, Pseudowire Setup and Maintenance Using the Label Distribution Protocol (LDP), April 2006 [RFC4448] Martini, L., and al, Encapsulation Methods for Transport of Ethernet over MPLS Networks, April 2006 [RFC4665] Augustyn & Serbest, Service Requirements for Layer 2 Provider-Provisioned Virtual Private Networks, September 2006 [RFC4761] Kompella & Rekhter, Virtual Private LAN Service (VPLS) Using BGP for Auto-Discovery and Signaling, January 2007 [RFC4762] Lasserre & Kompella, Virtual Private LAN Service (VPLS) Using Label Distribution Protocol (LDP) Signaling, January 2007 [RFC5501] Kamite, et al., Requirements for Multicast Support in Virtual Private LAN Services, March 2009 10.2. Informative References [Draft VPLS ETree Req] Key, et al., Requirements for MEF E-Tree Support in VPLS, draft-ietf-l2vpn-etree-reqt-00 (work in progress), October 2011 Key, et al. Expires April 2012 [Page 15]
Internet Draft Framework E-Tree over MPLS October 2011 [Draft CW L-bit] Delord, et al., Control Word Reserved bit for use in E-Tree, draft-delord-pwe3-cw-bit-etree-06 (work in progress), October 2011 [Draft VPLS ETree] Key, et al., Extension to VPLS for E-Tree, draft-key-l2vpn-vpls-etree-06 (work in progress), October 2011 [Draft VPLS PE ETree] Jiang, et al., VPLS PE Model for E-Tree Support, draft-jiang-l2vpn-vpls-pe-etree-04 (work in progress), July 2011 [Draft VPLS ETree 2PW] Ram, et al., Extension to LDP-VPLS for E-Tree Using Two PW, draft-ram-l2vpn-ldp-vpls-etree-2pw-02 (work in progress), May 2011 [Draft VPLS ETree 2VEID] Cao, Extension to Signaling in VPLS for E-Tree, draft-cao-l2vpn-vpls-etree-00 (work in progress), May 2011 [Draft P2MP PW Req] Jounay, et al., Requirements for Point-to-Multipoint Pseudowire, draft-ietf-pwe3-p2mp-pw-requirements-05 (work in progress), September 2011 [Draft P2MP PW Sig] Boutros, et al., Signaling Root-Initiated Point-to- Multipoint Pseudowires using LDP, draft-ietf-pwe3-p2mp-pw-02 (work in progress), March 2011 [Draft LDP-VPLS Bcast] Delord, et al., Extension to LDP-VPLS for Ethernet Broadcast and Multicast, draft-ietf-l2vpn-ldp-vpls-broadcast-exten-02 (work in progress), June 2011 [Draft Mcast VPLS] Raggarwa, Kamite & Fang, Multicast in VPLS, draft-ietf-l2vpn-vpls-mcast-09 (work in progress), July 2011 [Draft VPMS Frmwk] Kamite, et al., Framework and Requirements for Virtual Virtual Private Multicast Service (VPMS), draft-ietf-l2vpn-vpms-frmwk-requirements-04 (work in progress), July 2011 Key, et al. Expires April 2012 [Page 16]
Internet Draft Framework E-Tree over MPLS October 2011 Appendix A. Some Possible Ways for Leaf-to-Leaf Communication Restriction This appendix briefly describes the following approaches: - Single Root Only (A.1) - Only one PE has Roots (A.2) - Only one PE with both Root & Leaf - Backhaul Root (A.3) - Backhaul Leaf (A.4) - H-VPLS Root (A.5) - H-VPLS Leaf (A.6) - Separate PEs for Root and Leaf (A.7) - Separate VSI for Root and Leaf - Internal Connection (A.8) - External Connection (A.9) - Separate PWs for "From Root" traffic and "From Leaf" traffic (A.10) - "From Root" or "From Leaf" derived from source MAC address (A.11) - Static MAC address configuration for Root AC (A.12) Reference Model for Leaf-to-Leaf Communication Restriction <------------E-Tree------------> +---------+ +---------+ | PE1 | | PE2 | +----+ | +---+ | | +---+ | +----+ |CE01+----AC1----+--+ | | | | +--+----AC5----+CE05| +----+ (Root AC) | | V | | | | V | | (Root AC) +----+ +----+ | | | | | | | | +----+ |CE02+----AC2----+--+ | | Ethernet | | +--+----AC6----+CE06| +----+ (Root AC) | | S +--+-----PW-----+--+ S | | (Root AC) +----+ +----+ | | | | | | | | +----+ |CE03+----AC3----+--+ | | | | +--+----AC7----+CE07| +----+ (Leaf AC) | | I | | | | I | | (Leaf AC) +----+ +----+ | | | | | | | | +----+ |CE04+----AC4----+--+ | | | | +--+----AC8----+CE08| +----+ (Leaf AC) | +---+ | | +---+ | (Leaf AC) +----+ | | | | +---------+ +---------+ For the diagrams in this appendix, "L" indicates the particular AC or PW belonging to the PE local split horizon group specifically for Leaf-to-Leaf Communication Restriction. No communication is allowed between any two members of a split horizon group. Key, et al. Expires April 2012 [Page 17]
Internet Draft Framework E-Tree over MPLS October 2011 A.1. Single Root Only <------------E-Tree------------> +---------+ +---------+ | PE1 | | PE2 | +----+ | +---+ | | +---+ | |CE01+----AC1----+--+ | | | | | | +----+ (Root AC) | | V | | | | V | | | | | | | | | | | | | | Ethernet | | | | | | S +L-+-----PW-----+--+ S | | +----+ | | | | | | | | +----+ |CE03+----AC3----+-L+ | | | | +L-+----AC7----+CE07| +----+ (Leaf AC) | | I | | | | I | | (Leaf AC) +----+ +----+ | | | | | | | | +----+ |CE04+----AC4----+-L+ | | | | +L-+----AC8----+CE08| +----+ (Leaf AC) | +---+ | | +---+ | (Leaf AC) +----+ | | | | +---------+ +---------+ Concerns: - Not fulfil multi-Root requirement of generic MEF E-Tree service definition A.2. Only one PE has Roots <------------E-Tree------------> +---------+ +---------+ | PE1 | | PE2 | +----+ | +---+ | | +---+ | |CE01+----AC1----+--+ | | | | | | +----+ (Root AC) | | V | | | | V | | +----+ | | | | | | | | |CE02+----AC2----+--+ | | Ethernet | | | | +----+ (Root AC) | | S +L-+-----PW-----+--+ S | | +----+ | | | | | | | | +----+ |CE03+----AC3----+-L+ | | | | +L-+----AC7----+CE07| +----+ (Leaf AC) | | I | | | | I | | (Leaf AC) +----+ +----+ | | | | | | | | +----+ |CE04+----AC4----+-L+ | | | | +L-+----AC8----+CE08| +----+ (Leaf AC) | +---+ | | +---+ | (Leaf AC) +----+ | | | | +---------+ +---------+ Concerns: - Deployment constraint Key, et al. Expires April 2012 [Page 18]
Internet Draft Framework E-Tree over MPLS October 2011 A.3. Only one PE with both Root & Leaf - Backhaul Root +---AC5(Root AC)---------------------------+ | | | +-AC6(Root AC)----------------------+ | | | | | | | | | | | | | +---+-+---+ +---------+ | | | | | | | | | | +----+ | ++-++ | | +---+ | | +-+--+ |CE01+----AC1----+--+ | | | | | | | |CE05| +----+ (Root AC) | | V | | | | V | | | +----+ +----+ | | | | | | | | | +----+ |CE02+----AC2----+--+ | | Ethernet | | | | +--+CE06| +----+ (Root AC) | | S +L-+-----PW-----+--+ S | | +----+ +----+ | | | | | | | | +----+ |CE03+----AC3----+-L+ | | | | +L-+----AC7----+CE07| +----+ (Leaf AC) | | I | | | | I | | (Leaf AC) +----+ +----+ | | | | | | | | +----+ |CE04+----AC4----+-L+ | | | | +L-+----AC8----+CE08| +----+ (Leaf AC) | +---+ | | +---+ | (Leaf AC) +----+ | PE1 | | PE2 | +---------+ +---------+ <------------E-Tree------------> Concerns: - Deployment constraint - Long fibre path Key, et al. Expires April 2012 [Page 19]
Internet Draft Framework E-Tree over MPLS October 2011 A.4. Only one PE with both Root & Leaf - Backhaul Leaf <------------E-Tree------------> +---------+ +---------+ | PE1 | | PE2 | +----+ | +---+ | | +---+ | +----+ |CE01+----AC1----+--+ | | | | +--+----AC5----+CE05| +----+ (Root AC) | | V | | | | V | | (Root AC) +----+ +----+ | | | | | | | | +----+ |CE02+----AC2----+--+ | | Ethernet | | +--+----AC6----+CE06| +----+ (Root AC) | | S +--+-----PW-----+--+ S | | (Root AC) +----+ +----+ | | | | | | | | +----+ |CE03+----AC3----+-L+ | | | | | | +--+CE07| +----+ (Leaf AC) | | I | | | | I | | | +----+ +----+ | | | | | | | | | +----+ |CE04+----AC4----+-L+ | | | | | | | |CE08| +----+ (Leaf AC) | ++-++ | | +---+ | | +-+--+ | L L | | | | | +---+-+---+ +---------+ | | | | | | | | | | | | | | | +-AC7(Leaf AC)----------------------+ | | | +---AC8(Leaf AC)---------------------------+ Concerns: - Deployment constraint - Long fibre path Key, et al. Expires April 2012 [Page 20]
Internet Draft Framework E-Tree over MPLS October 2011 A.5. Only one PE with both Root & Leaf - H-VPLS Root <------------E-Tree------------> +---------+ +---------+ | PE1 | | PE2 | +----+ | +---+ | Ethernet | | +----+ |CE01+----AC1----+--+ +--+-----PW-----+---------+----AC5----+CE05| +----+ (Root AC) | | V | | | | (Root AC) +----+ +----+ | | | | Ethernet | | +----+ |CE02+----AC2----+--+ +--+-----PW-----+---------+----AC6----+CE06| +----+ (Root AC) | | S | | | +---+ | (Root AC) +----+ +----+ | | | | | | | | +----+ |CE03+----AC3----+-L+ | | Ethernet | | V +L-+----AC7----+CE07| +----+ (Leaf AC) | | I +L-+-----PW-----+--+ S | | (Leaf AC) +----+ +----+ | | | | | | I | | +----+ |CE04+----AC4----+-L+ | | | | +L-+----AC8----+CE08| +----+ (Leaf AC) | +---+ | | +---+ | (Leaf AC) +----+ | | | | +---------+ +---------+ Concerns: - Design complexity - More PW - Hair pinning (e.g. CE05 to CE06/07/08) impact bandwidth and delay A.6. Only one PE with both Root & Leaf - H-VPLS Leaf <------------E-Tree------------> +---------+ +---------+ | PE1 | | PE2 | +----+ | +---+ | | +---+ | +----+ |CE01+----AC1----+--+ | | | | +--+----AC5----+CE05| +----+ (Root AC) | | V | | Ethernet | | V | | (Root AC) +----+ +----+ | | +--+-----PW-----+--+ S | | +----+ |CE02+----AC2----+--+ | | | | I +--+----AC6----+CE06| +----+ (Root AC) | | S | | | | | | (Root AC) +----+ +----+ | | | | Ethernet | +---+ | +----+ |CE03+----AC3----+-L+ +L-+-----PW-----+---------+----AC7----+CE07| +----+ (Leaf AC) | | I | | | | (Leaf AC) +----+ +----+ | | | | Ethernet | | +----+ |CE04+----AC4----+-L+ +L-+-----PW-----+---------+----AC8----+CE08| +----+ (Leaf AC) | +---+ | | | (Leaf AC) +----+ | | | | +---------+ +---------+ Concerns: - Design complexity - More PW - Hair pinning (e.g. CE08 to CE05/06) impact bandwidth and delay Key, et al. Expires April 2012 [Page 21]
Internet Draft Framework E-Tree over MPLS October 2011 A.7. Separate PEs for Root and Leaf (PE2 split to PE2R & PE2L) <------------E-Tree------------> +---------+ +---------+ | PE1 | | PE2R | +----+ | +---+ | | +---+ | +----+ |CE01+----AC1----+--+ | | Ethernet | | V +--+----AC5----+CE05| +----+ (Root AC) | | V +--+-----PW-----+--+ S | | (Root AC) +----+ +----+ | | | | | | I | | +----+ |CE02+----AC2----+--+ | | | | +--+----AC6----+CE06| +----+ (Root AC) | | S | | | +-+-+ | (Root AC) +----+ +----+ | | | | | L | |CE03+----AC3----+-L+ | | +----+----+ +----+ (Leaf AC) | | I | | | +----+ | | | | |Ethernet |CE04+----AC4----+-L+ | | |PW +----+ (Leaf AC) | +-+-+ | | | L | +----+----+ +----+----+ | | | | | +-+-+ | +----+ | Ethernet | | V +L-+----AC7----+CE07| +----------PW-----+--+ S | | (Leaf AC) +----+ | | I | | +----+ | | +L-+----AC8----+CE08| | +---+ | (Leaf AC) +----+ | PE2L | +---------+ <------------E-Tree------------> Concerns: - Require two PEs in one POP - More PW Key, et al. Expires April 2012 [Page 22]
Internet Draft Framework E-Tree over MPLS October 2011 A.8. Separate VSI for Root and Leaf - Internal Connection (VSI on PE split to VSIR & VSIL) <------------E-Tree------------> +---------+ +---------+ | PE1 | | PE2 | +----+ | +---+ | | +---+ | +----+ |CE01+----AC1----+--+ V | | | | V +--+----AC5----+CE05| +----+ (Root AC) | | S +--+------------+--+ S | | (Root AC) +----+ +----+ | | I | | | | I | | +----+ |CE02+----AC2----+--+ R +L-+--+ +--+-L+ R +--+----AC6----+CE06| +----+ (Root AC) | +-+-+ | | | | +-+-+ | (Root AC) +----+ | L | | | | L | | | | \ / | | | | | | \ / | | | | | | \ / | | | Internal| | \/ | |Internal Connection| | /\ | |Connection | | | / \ | | | | | | / \ | | | | | | / \ | | | +----+ | +-+-+ | | | | +-+-+ | +----+ |CE03+----AC3----+-L+ V | | | | | | V +L-+----AC7----+CE07| +----+ (Leaf AC) | | S +--+--+ +--+--+ S | | (Leaf AC) +----+ +----+ | | I | | | | I | | +----+ |CE04+----AC4----+-L+ L | | Three | | L +L-+----AC8----+CE08| +----+ (Leaf AC) | +-+-+ | Ethernet | +---+ | (Leaf AC) +----+ | | PWs | | +---------+ +---------+ Concerns: - Design complexity - More VSI - More PW - Some vendor implementation may require additional hardware module to support internal connection between two VSIs - Some vendor implementation may have bandwidth limitation on internal connection between two VSIs - Some vendor implementation of service-aware management system may assume only one VSI per VPLS on one PE Key, et al. Expires April 2012 [Page 23]
Internet Draft Framework E-Tree over MPLS October 2011 A.9. Separate VSI for Root and Leaf - External Connection (VSI on PE split to VSIR & VSIL) <------------E-Tree------------> +---------+ +---------+ | PE1 | | PE2 | +----+ | +---+ | | +---+ | +----+ |CE01+----AC1----+--+ V | | | | V +--+----AC5----+CE05| +----+ (Root AC) | | S +--+------------+--+ S | | (Root AC) +----+ +----+ | | I | | | | I | | +----+ |CE02+----AC2----+--+ R +L-+--+ +--+-L+ R +--+----AC6----+CE06| +----+ (Root AC) | | | | | | | | | | (Root AC) +----+ | | | | | | | | | | +------AC-X1---+-L+ | | \ / | | +L-+----AC-X2-----+ | (Leaf AC) | +---+ | \ / | +---+ | (Leaf AC) | | | | \ / | | | |External | | \/ | | External| |Connection | | /\ | | Connection| | | +---+ | / \ | +---+ | | +------AC-Y1---+--+ | | / \ | | +--+----AC-Y2-----+ (Root AC) | | | | / \ | | | | (Root AC) +----+ | | | | | | | | | | +----+ |CE03+----AC3----+-L+ V | | | | | | V +L-+----AC7----+CE07| +----+ (Leaf AC) | | S +--+--+ +--+--+ S | | (Leaf AC) +----+ +----+ | | I | | | | I | | +----+ |CE04+----AC4----+-L+ L | | Three | | L +L-+----AC8----+CE08| +----+ (Leaf AC) | +-+-+ | Ethernet | +---+ | (Leaf AC) +----+ | | PWs | | +---------+ +---------+ Concerns: - Design complexity - More VSI - More PW - More AC (for external connection between two VSIs) - Require additional two high speed physical ports on PE to support such external connections - Some vendor implementation of service-aware management system may assume only one VSI per VPLS on one PE Key, et al. Expires April 2012 [Page 24]
Internet Draft Framework E-Tree over MPLS October 2011 A.10. Separate PWs for "From Root" traffic and "From Leaf" traffic <------------E-Tree------------> +---------+ +---------+ | PE1 | | PE2 | +----+ | +---+ | | +---+ | +----+ |CE01+----AC1----+--+ | | Ethernet | | +--+----AC5----+CE05| +----+ (Root AC) | | V +--+---PW for---+--+ V | | (Root AC) +----+ +----+ | | | |"From Root" | | | | +----+ |CE02+----AC2----+--+ | | Traffic | | +--+----AC6----+CE06| +----+ (Root AC) | | S | | | | S | | (Root AC) +----+ +----+ | | | | | | | | +----+ |CE03+----AC3----+--+ | | | | +--+----AC7----+CE07| +----+ (Leaf AC) | | I | | Ethernet | | I | | (Leaf AC) +----+ +----+ | | +--+---PW for---+--+ | | +----+ |CE04+----AC4----+--+ | |"From Leaf" | | +--+----AC8----+CE08| +----+ (Leaf AC) | +---+ | Traffic | +---+ | (Leaf AC) +----+ | | | | +---------+ +---------+ Concerns: - More PW - Most, if not all, vendor implementation support only one PW between two VSIs on different PEs - Most, if not all, vendor implementation of service-aware management system assume only one PW between two VSIs on different PEs - Asymmetric path for bidirectional traffic between Root and Leaf on different PEs (e.g. CE01-->CE07 use the "From Root" PW, CE07-->CE01 use the "From Leaf" PW) - Require extension to current standard VPLS - support two PWs between two VSIs on different PEs (both active but no loop) - share MAC learning between the "From Root" PW and "From Leaf" PW (bidirectional traffic may be on asymmetric path) - in addition to standard MAC-based forwarding, select which PW to use based on whether ingress AC is Root or Leaf - filter Leaf-to-Leaf traffic (split horizon group at PW/AC level is not good enough because of asymmetric path) Key, et al. Expires April 2012 [Page 25]
Internet Draft Framework E-Tree over MPLS October 2011 A.11. "From Root" or "From Leaf" derived from source MAC address Based on the current standard VPLS, a PE has no information about ACs on another PE. This approach will need additional information exchange between ingress PE and egress PE, via OSS or peer to peer. Concerns: - Require system development or additional signaling between PEs - Not an ideal solution from security perspective because of the dynamic nature of MAC address to AC mapping A.12. Static MAC address configuration for Root AC This approach requires additional configuration on PEs - Disable MAC address learning for Root ACs - Static configuration of MAC addresses per Root AC - Add filtering for each Root AC - Drop ingress frame if source MAC address not equal to any of the static MAC addresses configured for the particular Root AC - Add filtering for each Leaf AC - Drop ingress frame if source MAC address equal to any of the static MAC addresses configured for any Root ACs of the VPLS instance - Drop egress frame if source MAC address not equal to any of the static MAC addresses configured for any Root ACs of the VPLS instance Concerns: - No MAC address learning capability for Root ACs - Need resources for maintaining the static MAC address configuration per Root AC Key, et al. Expires April 2012 [Page 26]
Internet Draft Framework E-Tree over MPLS October 2011 Authors' Addresses Raymond Key Huawei Email: raymond.key@ieee.org Simon Delord Alcatel-Lucent Email: simon.delord@gmail.com Frederic Jounay Orange France Telecom 2, avenue Pierre-Marzin 22307 Lannion Cedex, France Email: frederic.jounay@orange.com Lucy Yong Huawei USA 1700 Alma Dr. Suite 500 Plano, TX 75075, USA Email: lucy.yong@huawei.com Lizhong Jin ZTE Corporation 889, Bibo Road Shanghai, 201203, China Email: lizhong.jin@zte.com.cn Yuji Kamite NTT Communications Corporation Granpark Tower 3-4-1 Shibaura, Minato-ku Tokyo 108-8118, Japan Email: y.kamite@ntt.com Wim Henderickx Alcatel-Lucent Copernicuslaan 50 2018 Antwerp, Belgium Email: wim.henderickx@alcatel-lucent.com Key, et al. Expires April 2012 [Page 27]
Internet Draft Framework E-Tree over MPLS October 2011 Copyright Notice Copyright (c) 2011 IETF Trust and the persons identified as the document authors. All rights reserved. This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (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. Key, et al. Expires April 2012 [Page 28]