Network Working Group B. Sarikaya Internet-Draft Huawei USA Intended status: Standards Track H. Ji Expires: April 2, 2015 China Telecom September 29, 2014 Multicast Support for Mapping of Address and Port Protocol and Light Weight 4over6 draft-sarikaya-softwire-map-multicast-03 Abstract This memo specifies multicast component for MAP-E and Light Weight 4over6 so that IPv4 hosts can receive multicast data from IPv4 servers over an IPv6 network. Two types of solutions are developed, encapsulation and translation. In the encapsulation solution for MAP-E and lw4o6, IGMP Proxy at the MAP-E Customer Edge router/lwB4 uses IPv4-in-IPv6 tunnel to exchange IGMP messages to establish multicast state at MAP-E Border Relay/lwAFTR so that MAP-E Border Relay/lwAFTR can tunnel IPv4 multicast data to IPv4 hosts. In the Translation Multicast solution for MAP-T, 4rd and lw4o6, IGMP messages are translated into MLD messages and sent to the network in IPv6. MAP-T/4rd Border Relay/lwAFTR does the reverse translation and joins IPv4 multicast group for the hosts. Border Relay/lwAFTR as multicast router receives IPv4 multicast data and translates the packet into IPv6 multicast data and sends downstream on the multicast tree. Member CEs/lwB4s receive multicast data, translate it back to IPv4 and transmit to the hosts. 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). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet- Drafts is at http://datatracker.ietf.org/drafts/current/. 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." This Internet-Draft will expire on April 2, 2015. Sarikaya & Ji Expires April 2, 2015 [Page 1] Internet-Draft Multicast Support for MAP-E and Lw4over6 September 2014 Copyright Notice Copyright (c) 2014 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. 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. Table of Contents 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 3 2. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 4 3. Architecture . . . . . . . . . . . . . . . . . . . . . . . . 4 3.1. Encapsulation Multicast Architecture . . . . . . . . . . 5 3.2. MAP-T and 4rd Translation Architecture . . . . . . . . . 7 4. MAP-E Encapsulation Multicast Operation . . . . . . . . . . . 9 4.1. Encapsulation Interface Considerations . . . . . . . . . 11 4.2. Avalanche Problem Considerations . . . . . . . . . . . . 12 5. lw4over6 Encapsulation Multicast Operation . . . . . . . . . 12 5.1. Tunnel Interface Considerations . . . . . . . . . . . . . 14 5.2. Dealing with Avalanche Problem . . . . . . . . . . . . . 15 5.3. Multicast Support for Host-Based Architecture . . . . . . 15 6. MAP-T and 4rd Translation Multicast Operation . . . . . . . . 15 6.1. Address Translation . . . . . . . . . . . . . . . . . . . 16 6.2. Protocol Translation . . . . . . . . . . . . . . . . . . 17 6.3. Learning Multicast Prefixes for IPv4-embedded IPv6 Multicast Addresses . . . . . . . . . . . . . . . . . . . 18 6.4. Supporting IPv4 Multicast at CE Router and lwB4 . . . . . 19 7. Security Considerations . . . . . . . . . . . . . . . . . . . 20 8. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 20 9. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 20 10. References . . . . . . . . . . . . . . . . . . . . . . . . . 20 10.1. Normative References . . . . . . . . . . . . . . . . . . 20 10.2. Informative references . . . . . . . . . . . . . . . . . 23 Appendix A. Group Membership Message Translation Details . . . . 23 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 24 Sarikaya & Ji Expires April 2, 2015 [Page 2] Internet-Draft Multicast Support for MAP-E and Lw4over6 September 2014 1. Introduction With IPv4 address depletion on the horizon, many techniques are being standardized for IPv6 migration including Mapping of Address and Port (MAP) - Encapsulation, - Translation and 4rd [I-D.ietf-softwire-map], [I-D.ietf-softwire-map-t], [I-D.ietf-softwire-4rd]. MAP/4rd enables IPv4 hosts to communicate with external hosts using IPv6 only ISP network. MAP/4rd Customer Edge (CE) device's LAN side is dual stack and WAN side is IPv6 only. CE tunnels/translates IPv4 packets received from the LAN side to 4rd Border Relays (BR). BRs have anycast IPv6 addresses and receive encapsulated/translated packets from CEs over a virtual interface. MAP/4rd operation is stateless. Packets are received/ sent independent of each other and no state needs to be maintained except for NAT44 operation on IPv4 packets received from the user. Light Weight 4 Over 6 (lw4o6) is a variant of Dual Stack Lite where carrier grade NAT is moved from AFTR to B4 element, i.e. NAPT is done locally at each B4 called light weight B4 or lwB4. Unicast lw4o6 takes user IPv4 packets from the local LAN and lwB4 does a NAPT and then tunnels the packets in an IPv4-in-IPv6 tunnel to lwAFTR which decapsulates the packet and then sends it to IPv4 network. Incoming packets follow reverse route and are encapsulated at lwAFTR and sent to lwB4 which decapsulates and after NAPT operation transmits to the destination. It should be noted that there is no depletion problem for IPv4 address space allocated for any source multicast and source specific multicast [RFC3171]. This document is not motivated by the depletion of IPv4 multicast addresses. MAP-E, MAP-T, 4rd and lw4o6 are unicast only. They do not support multicast. In this document we specify how multicast from home IPv4 users can be supported in MAP-E (as well as MAP-T and 4rd) and lw4o6. In case of MAP-E and lw4o6 tunneling multicast solution we integrate the multicast solution into the MAP-E/lw4o6 tunnel resulting in a multicast tunneling protocol. Multicast tunneling protocol has the advantage of not requiring multicast enabled IPv6 network between CE routers/lwB4s and MAP-E BRs/lwAFTRs. Similar tunneling is used in mobility protocols like Mobile IP [RFC6275] and Proxy Mobile IP [RFC6224]. In case IPv6 network is multicast enabled, MAP-T/4rd can provide multicast service to the hosts using MAP-T/4rd Multicast Translation based solution. A Multicast Translator can be used that receives IPv4 multicast group management messages in IGMP and generates corresponding IPv6 group management messages in MLD and sends them to Sarikaya & Ji Expires April 2, 2015 [Page 3] Internet-Draft Multicast Support for MAP-E and Lw4over6 September 2014 IPv6 network towards MAP-T/4rd Border Relay. We use [I-D.ietf-softwire-map-t] or [I-D.ietf-softwire-4rd] for sending IPv4 multicast data in IPv6 to the CE routers. At MAP-T/4rd CE router another translator is needed to translate IPv6 multicast data into IPv4 multicast data. It should be noted that if IPv6 network is multicast enabled the translation multicast solution presented in Section 6 can also be used for MAP-E. In this document we address MAP-E (and MAP-T/4rd) and lw4o6 multicast problem and propose two architectures: Multicast Tunneling and Multicast Translation based solutions. Section 2 is on terminology, Section 3 is on architecture, Section 4 and Section 5 are on multicast tunneling protocol Section 6 is on multicast translation protocol, and Section 7 states security considerations. 2. Terminology This document uses the terminology defined in [I-D.ietf-softwire-map], [I-D.ietf-softwire-lw4over6], [I-D.ietf-softwire-map-t], [I-D.ietf-softwire-4rd], [RFC3810] and [RFC3376]. 3. Architecture In MAP-E, MAP-T and 4rd, there are hosts (possibly IPv4/ IPv6 dual stack) served by MAP-E, MAP-T and 4rd Customer Edge device. CE is dual stack facing the hosts and IPv6 only facing the network or WAN side. MAP-E, MAP-T and 4rd CE may be local IPv4 Network Address and Port Translation (NAPT) box [RFC3022] by assigning private IPv4 addresses to the hosts. MAP-E, MAP-T and 4rd CEs in the same domain may use shared public IPv4 addresses on their WAN side and if they do they should avoid ports outside of the allocated port set for NAPT operation. At the boundary of the network there is MAP-E, MAP-T and 4rd Border Relay. BR receives IPv4 packets tunneled in IPv6 from CE and decapsulates them and sends them out to IPv4 network. Unicast MAP-E, MAP-T and 4rd are stateless except for the local NAPT at the CE. Each IPv4 packet sent by CE treated separately and different packets from the same CE may go to different BRs or CEs. CE encapsulates IPv4 packet in IPv6 with destination address set to BR address (usually anycast IPv6 address). BR receives the encapsulated packet and decapsulates and send it to IPv4 network. CEs are configured with Rule IPv4 Prefixes, Rule IPv6 Prefixes and with an BR IPv6 anycast address. BR receives IPv4 packets addressed to this ISP and from the destination address it extracts the destination host's IPv4 address and uses this address as destination Sarikaya & Ji Expires April 2, 2015 [Page 4] Internet-Draft Multicast Support for MAP-E and Lw4over6 September 2014 address and encapsulates the IPv4 packet in IPv6 and sends it to IPv6-only network. Unicast Lightweight 4over6 (lw4o6) is a variation of Dual-Stack Lite (DS-Lite) [RFC6333] which moves carrier-grade IPv4-IPv4 NAT from the Address Family Transition Router (AFTR) element to the Basic Bridging BroadBand (B4) element [I-D.ietf-softwire-lw4over6]. The resulting elements are called lwAFTR and lwB4 with NAPT, respectively. Lw4o6 also adopts some features from MAP-E. A+P scheme of public IPv4 address sharing is used by lwB4's in assigning WAN side IPv4 public addresses with a distinct port set. As in MAP-E, encapsulation of IPv4 packets in IPv6 and decapsulation is according to [RFC2473]. 3.1. Encapsulation Multicast Architecture Encapsulation variant of MAP called MAP-E network lends itself easily to the Multicast Tunneling architecture. Dual stack hosts are connected to the Customer Edge router and it is multicast enabled. It is assumed that IPv6 only network is unicast only network and that IPv6 multicast is not enabled or IPv6 multicast is partially enabled, e.g. the access network is not enabled for IPv6 multicast. At the boundary of the network MAP-E Border Relay is connected to the native multicast backbone infrastructure. We place IGMP Proxy at the CE router. CE router serves all the connected hosts. For multicast traffic, CE Router uses MAP-E tunneling interface with MAP-E BR to send/receive IGMP messages using IPv4-in-IPv6 tunnel [RFC2473]. MAP-E BR is IGMP Router towards the CEs and it could be IGMP Router or PIM router upstream. A given relay and all CEs connected to it can be considered to be on a separate logical link. On this link, gateways and relay communicate using IPv4-in-IPv6 tunneling to transmit and receive multicast control messages for membership management and multicast data from the relay to the gateways. All the elements of MAP-E multicast support system with tunneling are shown in Figure 1. Sarikaya & Ji Expires April 2, 2015 [Page 5] Internet-Draft Multicast Support for MAP-E and Lw4over6 September 2014 Dual Stack Hosts IPv4 +----+ Network | H1 | IPv6 +----+ +-------+ only +--------------+ + +----+ | CE | network | BR | | H2 | ---| IGMP |--- -- |IGMP | IGMP | IPv6 +----+ | Proxy | |Router| PIM | Network +----+ +-------+ +--------------+ | H3 | +----+ Figure 1: Architecture of MAP-E Multicast Tunneling For lw4o6 multicast support with tunneling, the architecture shown in Figure 2 will be used. lwB4 is IGMP Proxy serving multicast receivers in IPv4 LAN. As such, lw4o6 B4 supporting multicast MUST conform IGMP Proxying protocol described in [RFC4605]. lwB4 uses lw4o6 tunneling interface with lwAFTR to send/receive IGMP messages using IPv4-in-IPv6 tunnel [RFC2473]. Due to the presence of NAPT at the lw4o6, lwB4 MUST conform to IP multicast requirements stated in [RFC5135]. lwAFTR is IGMP Router towards the l4B4s and it could be IGMP Router or PIM router upstream. A given relay and all lwB4s connected to it can be considered to be on a separate logical link. On this link, lwB4 and lwAFTR communicate using IPv4-in-IPv6 tunneling to transmit and receive multicast control messages for membership management and multicast data from lwAFTR to the lwB4s. IPv4 +-------+ IPv4 Network | |Multicast | | +-------+ +--------------+ + | | | lwB4 | IPv4-in-IPv6 | lwAFTR | | IPv4 |----->| IGMP |-------------->|IGMP | IGMP | IPv6 | LAN |<-----| Proxy |<--------------|Router| PIM | Network | | +-------+ +--------------+ | | +-------+ Figure 2: Architecture of lw4o6 Multicast Tunneling Sarikaya & Ji Expires April 2, 2015 [Page 6] Internet-Draft Multicast Support for MAP-E and Lw4over6 September 2014 3.2. MAP-T and 4rd Translation Architecture In case IPv6 only network is multicast enabled, translation multicast architecture can be used. CE implements IGMP Proxy function [RFC4605] towards the LAN side and MLD Proxy on its WAN interface. IPv4 hosts send their join requests (IGMP Membership Report messages) to CE. CE as a MLD proxy sends aggregated MLD Report messages upstream towards BR. CE replies MLD membership query messages with MLD membership report messages based on IGMP membership state in the IGMP/MLD proxy. BR is MLD querier on its WAN side. On its interface to IPv4 network BR may either have IGMP client or PIM. PIM being able to support both IPv4 and IPv6 multicast should be preferred. BR receives MLD join requests, extracts IPv4 multicast group address and then joins the group upstream, possibly by issuing a PIM join message. IPv4 multicast data received by the BR as a leaf node in IPv4 multicast distribution tree is translated into IPv6 multicast data by the translator using [I-D.ietf-softwire-map-t], [I-D.ietf-softwire-4rd] and then sent downstream to the IPv6 part of the multicast tree to all downstream routers that are members. IPv6 data packet eventually gets to the CE. At the CE, a reverse [I-D.ietf-softwire-map-t], [I-D.ietf-softwire-4rd] operation takes place by the translator and then IPv4 multicast data packet is sent to the member hosts on the LAN interface. [I-D.ietf-softwire-map-t], [I-D.ietf-softwire-4rd] are modified to handle multicast addresses. In order to support SSM, IGMPv3 MUST be supported by the host, CE and BR. For ASM, BR MUST be the Rendezvous Point (RP). MAP-T and 4rd Translation Multicast solution uses the multicast 46 translator in not one but two places in the architecture: at the CE router and at the Border Relay. IPv4 multicast data received at 4rd BR goes through a [I-D.ietf-softwire-4rd] header-mapping into IPv6 multicast data at the BR and another [I-D.ietf-softwire-4rd] header- mapping back to IPv4 multicast data at the CE router. Encapsulation variant of [I-D.ietf-softwire-4rd] is not used. In case of MAP-T, IPv4 data packet is translated using v4 to v6 header translation using multicast addresses instead of the mapping algorithm used in [I-D.ietf-softwire-map-t]. All the elements of MAP-T and 4rd translation-based multicast support system are shown in Figure 3. Sarikaya & Ji Expires April 2, 2015 [Page 7] Internet-Draft Multicast Support for MAP-E and Lw4over6 September 2014 Dual Stack Hosts IPv4 +----+ Network | H1 | +----------+ IPv6 +-------------+ | | | CE | | BR | +----+ |Translator| only | Translator | |MAP-T/4rd | | MAP-T/4rd | +----+ | | network | |IGMP| + | H2 | ---|IGMP-MLD |--------- -- |MLD | or | IPv6 +----+ | Proxy | |Querier |PIM | Network +----+ +----------+ +-------------+ | H3 | +----+ Figure 3: Architecture of MAP-T and 4rd Translation Multicast In case IPv6 only network is multicast enabled, translation multicast architecture can also be used for lw4o6 multicast. lwB4 implements IGMP Proxy function [RFC4605] towards the LAN side and MLD Proxy on its WAN interface. IPv4 hosts send their join requests (IGMP Membership Report messages) to lwB4. lwB4 as a MLD proxy sends aggregated MLD Report messages upstream towards lwAFTR. lwB4 replies MLD membership query messages with MLD membership report messages based on IGMP membership state in the IGMP/MLD proxy. lwAFTR is MLD querier on its WAN side. On its interface to IPv4 network lwAFTR may either have IGMP client or PIM. PIM being able to support both IPv4 and IPv6 multicast should be preferred. lwAFTR receives MLD join requests, extracts IPv4 multicast group address and then joins the group upstream, possibly by issuing a PIM join message. For multicast data, [I-D.ietf-softwire-dslite-multicast] uses encapsulation of IPv4 multicast data in IPv6 multicast data packet but in this document we use translation. IPv4 multicast data received by the lwAFTR as a leaf node in IPv4 multicast distribution tree is translated into IPv6 multicast data by the translator and then sent downstream to the IPv6 part of the multicast tree to all downstream routers that are members. IPv6 data packet eventually gets to the lwB4. At the lwB4, a reverse translation operation takes place by the translator and then IPv4 multicast data packet is sent to the member hosts on the LAN interface. The translation algorithm in [I-D.ietf-softwire-map-t], [I-D.ietf-softwire-4rd] are modified to handle multicast addresses. In order to support SSM, IGMPv3 MUST be supported by the host, lwB4 and lwAFTR. For ASM, lwAFTR MUST be the Rendezvous Point (RP). Sarikaya & Ji Expires April 2, 2015 [Page 8] Internet-Draft Multicast Support for MAP-E and Lw4over6 September 2014 MAP-T and 4rd Translation Multicast solution uses the multicast 46 translator in not one but two places in the architecture: at the lwB4 router and at the lwAFTR. IPv4 data packet is translated using v4 to v6 header translation using multicast addresses instead of the mapping algorithm used in [I-D.ietf-softwire-map-t]. All the elements of lw4o6 translation-based multicast support system are shown in Figure 4. IPv4 +-------+ IPv4 Network | |Multicast | | +--------------+ +--------------+ + | | | lwB4 NAPT |IPv6 | lwAFTR | | IPv4 |----->| IGMP MLD |----->|IGMP | IGMP | IPv6 | LAN |<-----| Proxy Proxy |------|Router| PIM | Network | | +--------------+ +--------------+ | | +-------+ Figure 4: Architecture of lw4o6 Multicast Translation 4. MAP-E Encapsulation Multicast Operation When a host (H1, H2 or H3 in Figure 1) wants to join an IPv4 multicast group G or (S,G), it sends an IGMP report (IGMPv3 report for a source-specific group) to CE router. CE encapsulates IGMP report messages in IPv6 and sends it over the tunnel to BR in anycast. CE router uses BR's anycast address this CE router is configured with. After CE receives unicast address of BR, it sends all subsequent IGMP messages for G or (S,G) in unicast. BR (topologically closest to this CE router) receives the message, decapsulates it and then lets IGMP router to process it. On the upstream, an IGMP Join message is sent to subscribe group G or (S,G) or a PIMv4 Join message is sent if PIM is supported. BR establishes membership state for group G or (S,G). BR sends all related IGMP messages to this CE in unicast using IPv4-in-IPv6 tunneling. CE now has BR's unicast address which it uses to send all IGMP packets for group G for any source multicast or (S,G) for source specific multicast. If CE receives multiple join messages for the same group G, CE sends an aggregated join message to BR. Sarikaya & Ji Expires April 2, 2015 [Page 9] Internet-Draft Multicast Support for MAP-E and Lw4over6 September 2014 If CE receives another join message for a different group G', (S',G') CE encapsulates it and sends it in anycast to the BR. This enables the use of multiple BRs that may be deployed as anchor points and makes downstream multicast data delivery more efficient. A CE is required to assist in IGMP signaling and data forwarding between the hosts that it serves and the corresponding BRs that are handling the multicast group G or (S,G). CE must have IGMP Proxy for each upstream tunnel interface that has been established with the BR. The CE decides on the mapping of downstream links to a proxy instance connected to an upstream link to a BR based on the unicast source IPv6 address in the packets received from BR. Because of this BRs MUST use the unicast source IPv6 address in packets sent to CEs. Encapsulation at the CE is according to [RFC2473] with an IPv4 payload carrying IGMP messages. On the reception of IGMP reports from the hosts, the CE must identify the corresponding proxy instance from the incoming interface and perform regular IGMP proxy operations of inserting, updating or removing multicast forwarding state on the incoming interface and will merge state updates into the IGMP proxy membership database. It will then send an aggregated Report via the upstream tunnel to the BR when the membership database changes. On the reception of IGMP queries, the CE proxy instance will answer the Queries on behalf of all active downstream receivers maintained in its membership database. Queries sent by the BR do not force the CE to trigger corresponding messages immediately towards hosts. BR acts as the default multicast querier for the corresponding CE. It implements the function of the designated multicast router or a further IGMP proxy. After BR receives IGMP Join message it adds the tunnel to the CE in its outgoing interface list for the group (G) or the source, group (S,G) that the host wants to join. BR establishes group-/source-specific multicast forwarding states at its corresponding downstream tunnel interfaces. Afterwards, BR maintains/removes these group-/source-specific multicast forwarding states. BR treats its tunnel interfaces as multicast-enabled downstream links, serving zero to many listening nodes. BR will send a join message upstream towards the source of the multicast group to build a multicast tree in the native multicast infrastructure and becomes a leaf node in the multicast tree. BR will send any group management messages (IGMP Report or Query messages) downstream to specific CEs on the tunnel interface by encapsulating these IGMP messages in IPv6 using [RFC2473]. Sarikaya & Ji Expires April 2, 2015 [Page 10] Internet-Draft Multicast Support for MAP-E and Lw4over6 September 2014 As for multicast data, the data packets from the source received at the BR will be replicated to all interfaces in it's outgoing interface list as well as the tunnel outgoing interface for all member CEs. BR sends multicast data in IPv4-in-IPv6 tunnel to the CE with the data packet encapsulated. Encapsulation is according to [RFC2473] with an IPv4 payload. CE receives Multicast Data message over the tunnel interface associated with the tunnel to BR. After decapsulation, multicast traffic arriving at the CE on an upstream interface will be forwarded according to the group-specific or source-specific forwarding states as acquired for each downstream interface within the IGMP proxy instance. 4.1. Encapsulation Interface Considerations Legacy IPv4 in IPv6 tunneling is performed as in [RFC2473]. Packets upstream from CE carry only IGMP signaling messages and they are not expected to be subject to fragmentation. However packets downstream, i.e. multicast data to CE may be subject to fragmentation. Source and destination addresses of IGMP messages in IPv4-in-IPv6 softwire from CE is as follows: Source address of IPv6 header is CE IPv6 address, e.g. 2001:db8:0:1::1, destination address is BR anycast address, possibly shared of the MAP domain. Source address of IGMP messages is CE's IPv4 interface address, e.g. 192.0.0.2, destination address is the all-systems multicast address of 224.0.0.1 for IGMP Query, all IGMPv3-capable multicast routers of 224.0.0.22 for IGMPv3 Report, the multicast group specified in the Group Address field of the Report for IGMPv1 or IGMPv2 Report. Source and destination addresses of IGMP messages in IPv4-in-IPv6 softwire from BR is as follows: Source address of IPv6 header is BR's unicast IPv6 address, e.g. 2001:db8:0:2::1, destination address is CE IPv6 address, e.g. 2001:db8:0:1::1. Source address of IGMP messages is CE's IPv4 interface address, e.g. 192.0.2.1, destination address is the all-systems multicast address of 224.0.0.1 for IGMP Query, all IGMPv3-capable multicast routers of 224.0.0.22 for IGMPv3 Report, the multicast group specified in the Group Address field of the Report for IGMPv1 or IGMPv2 Report. Sarikaya & Ji Expires April 2, 2015 [Page 11] Internet-Draft Multicast Support for MAP-E and Lw4over6 September 2014 Source and destination addresses of multicast data messages in IPv4- in-IPv6 softwire is as follows: Source address of IPv6 header is BR IPv6 unicast address, e.g. 2001:db8:0:2::1, destination address is CE IPv6 address, e.g. 2001:db8:0:1::1. Source address of IPv4 multicast data is unicast IPv4 address of the multicast source, e.g. the content provider, destination address is IPv4 multicast group address. BR decapsulates datagrams carrying IGMP messages from CE's and then IGMP/PIM router processing takes over. Network Address Translation (NAT) is not applied on IGMP messages. 4.2. Avalanche Problem Considerations In Section 4 BR replicates the data packets from the source received to all outgoing interfaces for all member CEs. This replication (often called avalanche problem) can be very costly if there are very large number of downstream member CEs such as in IPTV application. Note that the avalanche problem is faced by all multicast solutions that use tunneling to bypass non-multicast enabled access network. In multicast MAP-E, one approach that can be used is to deploy MAP-E BRs close to the user. BRs colocated at the access network gateway such as at the Border Network Gateway (BNG) could reduce the packet duplication bottleneck considerably. In multicast MAP-E, another approach is to exploit multiple BRs that can be deployed in the network. MAP-E CE can use BR anycast address when sending an encapsulated upstream IGMP join request and then use the unicast source address of this BR in subsequent IGMP messages. 5. lw4over6 Encapsulation Multicast Operation The hosts will send their subscription requests for IPv4 multicast groups upstream to the default router, i.e. lwB4 Element. After subscribing to the group, the host can receive multicast data from the lwB4. The host implements IGMP protocol's host part. In order to support SSM, IGMPv3 MUST be supported by the host, l4B4 and lwAFTR. lwB4 Element is IGMP Proxy. After receiving the first IGMP Report message requesting subscription to an IPv4 multicast group, lwB4 establishes a tunnel interface with a lwAFTR. The tunnel is IPv6 based but it will carry IPv4 traffic, IGMP messages back and forth Sarikaya & Ji Expires April 2, 2015 [Page 12] Internet-Draft Multicast Support for MAP-E and Lw4over6 September 2014 and IPv4 multicast data messages downstream. lwB4 does not have to keep more than one tunnel interfaces, a single interface is sufficient. IGMP Proxy at the lwB4 does not have to have more than one proxy instances, a single instance is sufficient. lwB4 is regular IGMP proxy and it keeps IGMP proxy membership database. lwB4 inserts multicast forwarding state on the incoming interface, and merges state updates into the IGMP proxy membership database. lwB4 updates or removes elements from the database as required. lwB4 will then send an aggregated Report via the upstream tunnel to the lwAFTR when the membership database changes. lwB4 answers IGMP queries from lwAFTR based on the membership database. lwB4's downstream link follows the traditional multipoint channel forwarding and does not pose any specific problems. lwB4 receives IPv4 multicast data from the lwAFTR tunneled over the tunnel interface. lwB4 decapsulates the packet and then forwards it downstream. Each member host receives the data packet based on Layer 2 multicast interface. No packet duplication is necessary. lwAFTR acts as the as the default multicast querier for all lwB4s that have established an IPv6 tunnel with it. In order to keep a consistent multicast state between a lwB4 and lwAFTR, once a lwB4 is connected it will stay connected until the state becomes empty. After that point, the lwB4 may continue to use the tunnel for IPv4 unicast traffic. According to aggregated IGMP reports received from a lwB4, lwAFTR establishes group/source-specific multicast forwarding states at its corresponding downstream tunnel interfaces. After that, lwAFTR maintains or removes the state as required by the aggregated reports received from lwB4. At the upstream interface, lwAFTR procures for aggregated multicast membership maintenance. Based on the multicast-transparent operations of the lwB4s, the lwAFTR treats its tunnel interfaces as multicast enabled downstream links, serving zero to many listening nodes. Multicast traffic arriving at the lwAFTR is transparently forwarded according to its multicast forwarding information base. Multicast data is first replicated and then forwarded in IPv4-in-IPv6 tunnel from lwAFTR to the corresponding lwB4s. Sarikaya & Ji Expires April 2, 2015 [Page 13] Internet-Draft Multicast Support for MAP-E and Lw4over6 September 2014 5.1. Tunnel Interface Considerations Legacy IPv4 in IPv6 tunneling is performed as in [RFC2473] and [RFC4213]. Considerations specified in [RFC6333] apply. Packets upstream from lwB4 carry only IGMP signaling messages and they are not expected to fragmentation. However packets downstream, i.e. multicast data to B4 may be subject to fragmentation. Source and destination addresses of IGMP messages in IPv4-in-IPv6 softwire from lwB4 is as follows: Source address of IPv6 header is lwB4 IPv6 address, e.g. 2001:db8:0:1::1, destination address is lwAFTR address, e.g. 2001:db8:0:2::1. Source address of IGMP messages is lwB4's IPv4 interface address, e.g. 192.0.0.2, destination address is the all-systems multicast address of 224.0.0.1 for IGMP Query, all IGMPv3-capable multicast routers of 224.0.0.22 for IGMPv3 Report, the multicast group specified in the Group Address field of the Report for IGMPv1 or IGMPv2 Report. Source and destination addresses of IGMP messages in IPv4-in-IPv6 softwire from lwAFTR is as follows: Source address of IPv6 header is lwAFTR address, e.g. 2001:db8:0:2::1, destination address is B4 IPv6 address, e.g. 2001:db8:0:1::1. Source address of IGMP messages is lwAFTR's IPv4 interface address, e.g. 192.0.2.1, destination address is the all-systems multicast address of 224.0.0.1 for IGMP Query, all IGMPv3-capable multicast routers of 224.0.0.22 for IGMPv3 Report, the multicast group specified in the Group Address field of the Report for IGMPv1 or IGMPv2 Report. Source and destination addresses of multicast data messages in IPv4- in-IPv6 softwire is as follows: Source address of IPv6 header is lwAFTR address, e.g. 2001:db8:0:2::1, destination address is lwB4 IPv6 address, e.g. 2001:db8:0:1::1. Source address of IPv4 multicast data is unicast IPv4 address of the multicast source, e.g. the content provider, destination address is IPv4 multicast group address. Sarikaya & Ji Expires April 2, 2015 [Page 14] Internet-Draft Multicast Support for MAP-E and Lw4over6 September 2014 lwAFTR decapsulates datagrams carrying IGMP messages from lwB4's and then IGMP router processing takes over. Network Address Translation (NAT) is not applied on IGMP messages. 5.2. Dealing with Avalanche Problem When multicast datagrams are received at the lwAFTR, lwAFTR consults its memebership database and duplicates the packets for each member lwB4 interface and then these datagrams are forwarded in IPv4-in-IPv6 softwire downstream. This may cause an avalanche of downstream packets if the number of member lwB4's is high. Avalanche problem can be eased by network partitioning. lwAFTR can be deployed closer to the users. For example in broadband networks, lwAFTR can be deployed at the Broadband Network Gateway (BNG) nodes. 5.3. Multicast Support for Host-Based Architecture In this section we specify multicast support for Host-Based DS-Lite architecture described in [I-D.ietf-softwire-lw4over6]. The host accesses the service provider network directly with an IPv6 global address. Host sends its IPv4 datagrams in IPv6 using an IPv4- in-IPv6 softwire tunnel to an lwAFTR, i.e. it implements lwB4. Source address of all IPv4 datagrams is the pre-configured well-known IPv4 non-routable address. For multicast, there are two choices: the host could implement host side of IGMP protocol or for mobile router type of hosts, the host implements IGMP proxy. Host encapsulates IGMP messages and sends them to lwAFTR. lwAFTR does not perform IPv4-IPv4 NAT translations on IGMP datagrams instead they are processed by IGMP router at the lwAFTR. Multicast data received from lwAFTR for a multicast group that the host has subscribed is decapsulated by the host, if the host is IGMP client, it processes the data. If the host is IGMP proxy, it consults multicast state for the group and forwards the data downstream so that the members can receive the data. 6. MAP-T and 4rd Translation Multicast Operation In this section we specify how the host can subscribe and receive IPv4 multicast data from IPv4 content providers based on the architecture defined in Figure 3 in two parts: address translation and protocol translation. Translation details are given in Appendix A. Sarikaya & Ji Expires April 2, 2015 [Page 15] Internet-Draft Multicast Support for MAP-E and Lw4over6 September 2014 6.1. Address Translation IPv4-only host, H1 will join IPv4 multicast group by sending IGMP Membership Report message upstream towards the IGMP Proxy in Figure 3. MLD Proxy first creates a synthesized IPv6 address of IPv4 multicast group address using IPv4-embedded IPv6 multicast address format [I-D.ietf-mboned-64-multicast-address-format]. ASM_MPREFIX64 for any source multicast groups and SSM_MPREFIX64 for source specific multicast groups are used. Both are /96 prefixes. SSM_MPREFIX64 is set to ff3x:0:8000::/96, with 'x' set to any valid scope. ASM_MPREFIX64 values are formed as shown in Figure 5. Flag field 1 (ff1) field is defined in [RFC7371] bits M bit MUST BE set to 1. "scop" field is defined in [RFC3956]. Flag field 2 (ff2) is a set of 4 flags rrrM where r bits MUST be set to zero. M bit is set to 1 to indicate that a multicast IPv4 address is embedded in the low- order 32 bits of the multicast IPv6 address. "sub-group-id" field MUST follow the recommendations specified in [RFC3306] if unicast- based prefix is used or the recommendations specified in [RFC3956] if embedded-RP is used. The default value is all zeros. | 8 | 4 | 4 | 4 | 76 | 32 | +--------+----+----+----+------------------------------+----------+ |11111111|ff1 |scop|ff2 | sub-group-id |v4 address| +--------+----+----+----+-----------------------------------------+ Figure 5: ASM_MPREFIX64 Formation Each translator in the upstream BR is assigned a unique ASM_MPREFIX64 prefix. CE (MLD Proxy in CE) can learn this value by means out of scope with this document. With this, CE can easily create an IPv6 multicast address from the IPv4 group address a.b.c.d that the host wants to join. Source-Specific Multicast (SSM) can also be supported similar to the Any Source Multicast (ASM) described above. In case of SSM, IPv4 multicast addresses use 232.0.0.0/8 prefix. IPv6 SSM_MPREFIX64 is set to FF3x:0:8000::/96 where 'x' is any valid scope. Since SSM translation requires a unique address for each IPv4 multicast source, an IPv6 unicast prefix must be configured to the translator in the upstream BR to represent IPv4 sources. This prefix is prepended to IPv4 source addresses in translated packets. The join message from the host for the group ASM_MPREFIX64:a.b.c.d or SSM_MPREFIX64:a.b.c.d or an aggregate join message will be received by MLD querier at the BR. BR as multicast anchor checks the group address and recognizes ASM_MPREFIX64 or SSM_MPREFIX64 prefix. It Sarikaya & Ji Expires April 2, 2015 [Page 16] Internet-Draft Multicast Support for MAP-E and Lw4over6 September 2014 next checks the last 32 bits is an IPv4 multicast address in range 224/8 - 239/8. If all checks succeed, IGMPv4 Client joins a.b.c.d using IGMP on its IPv4 interface. Joining IPv4 groups can also be done using PIM since PIM supports both IPv4 and IPv6. The advantage of using PIM is that there is no need to enable IGMP support in neighboring IPv4 routers. The advantage of using IGMP is that IGMP is a simpler protocol and it is supported by a wider range of routers. The use of PIM or IGMP is left as an implementation choice. Address translation described above for MAP-T applies to lw4over6 multicast translation where the entities involved are lwB4 replaces Customer Edge device and lwAFTR replaces BR Figure 4. 6.2. Protocol Translation The hosts will send their subscription requests for IPv4 multicast groups upstream to the default router, i.e. Costumer Edge device. After subscribing the group, the host can receive multicast data from the CE. The host implements IGMP protocol's host part. Customer Edge device is IGMP Proxy facing the LAN interface. After receiving the first IGMP Report message requesting subscription to an IPv4 multicast group, a.b.c.d, MLD Proxy in the CE's WAN interface synthesizes an IPv6 multicast group address corresponding to a.b.c.d and sends an MLD Report message upstream to join the group. When MAP-T or 4rd BR receives IPv4 multicast data for an IPv4 group a.b.c.d it [I-D.ietf-softwire-4rd] translates/encapsulates IPv4 packet into IPv6 multicast packet and sends it to IPv6 synthesized address corresponding to a.b.c.d using ASM_MPREFIX64 or SSM_MPREFIX64. The header mapping described in [I-D.ietf-softwire-4rd] Section 4.2 (using Table 1) is used except for mapping the source and destination addresses. In this document we use the multicast address translation described in Section 6.1 and propose it as a complementary enhancement to the translation algorithm in [I-D.ietf-softwire-4rd]. The IP/ICMP translation translates IPv4 packets into IPv6 using minimum MTU size of 1280 bytes (Section 4.3 in [I-D.ietf-softwire-4rd]) but this can be changed for multicast. Path MTU discovery for multicast is possible in IPv6 so 4rd BR can perform path MTU discovery for each ASM group and use these values instead of 1280. For SSM, a different MTU value MUST be kept for each SSM channel. Because of this 8 bytes added by IPv6 fragment header in each data packet can be tolerated. Sarikaya & Ji Expires April 2, 2015 [Page 17] Internet-Draft Multicast Support for MAP-E and Lw4over6 September 2014 Since multicast address translation does not preserve checksum neutrality, [I-D.ietf-softwire-4rd] translator/encapsulator at 4rd BR must however modify the UDP checksum to replace the IPv4 addresses with the IPv6 source and destination addresses in the pseudo-header which consists of source address, destination address, protocol and UDP length fields before calculating the new checksum. IPv6 multicast data must be translated back to IPv4 at the 4rd CE (e.g. using Table 2 in Section 4.3 of [I-D.ietf-softwire-4rd]). Such a task is much simpler than the translation at 4rd BR because IPv6 header is much simpler than IPv4 header and IPv4 link on the LAN side of 4rd CE is a local link. The packet is sent on the local link to IPv4 group address a.b.c.d for IPv6 group address of ASM_MPREFIX64:a.b.c.d or SSM_MPREFIX64:a.b.c.d. In case an IPv4 multicast source sends multicast data with the don't fragment (DF) flag set to 1, [I-D.ietf-softwire-4rd] header mapping sets the D bit in IPv6 fragment header before sending the packet downstream as in Fig. 3 in Section 4.3 of [I-D.ietf-softwire-4rd]. This feature of [I-D.ietf-softwire-4rd] preserves the semantics of DF flag at the BR and CE. Because MAP-T/4rd is stateless, Multicast MAP-T/4rd should stay faithful to this as much as possible. Border Relay acts as the default multicast querier for all CEs that have established multicast communication with it. In order to keep a consistent multicast state between a CE and BR, CE MUST use the same IPv6 multicast prefixes (ASM_MPREFIX64/SSM_REFIX64) until the state becomes empty. After that point, the CE may obtain different values for these prefixes, effectively changing to a different 4rd BR. Protocol translation described above for MAP-T applies to lw4over6 multicast translation where the entities involved are lwB4 replaces Customer Edge device and lwAFTR replaces BR Figure 4. 6.3. Learning Multicast Prefixes for IPv4-embedded IPv6 Multicast Addresses CE can be pre-configured with Multicast Prefix64 of ASM_MPREFIX64 and SSM_MPREFIX64 that are supported in their network. However automating this process is also desired. A new router advertisement option, a Multicast ASM Translation Prefix option, can be defined for this purpose. The option contains IPv6 ASM multicast translation prefix, ASM_MPREFIX64. A new router advertisement option, a Multicast SSM Translation Prefix option, can be defined for this purpose. The option contains IPv6 SSM multicast prefix translation prefix SSM_MPREFIX64. Sarikaya & Ji Expires April 2, 2015 [Page 18] Internet-Draft Multicast Support for MAP-E and Lw4over6 September 2014 After the host gets the multicast prefixes, when an application in the host wishes to join an IPv4 multicast group the host MUST use ASM_MPREFIX64 or SSM_MPREFIX64 and then obtain the synthesized IPv6 group address before sending MLD join message. Source-specific multicast (SSM) group membership message payloads in IGMPv3 and MLDv2 contain address literals and their translation requires another multicast translation prefix option. IPv4 source addresses in IGMPv3 Membership Report message are unicast addresses of IPv4 sources and they have to be translated into unicast IPv6 source addresses in MLDv2 Membership Report message. A new router advertisement option, a Multicast Translation Unicast Prefix option can be defined for this purpose. The option contains IPv6 unicast Network-Specific Prefix U_PREFIX64. The host can be configured by its default router using router advertisements containing the prefixes [I-D.sarikaya-softwire-6man-raoptions]. 64:ff9b::/96 is the global value called well-known prefix that is assigned to U_PREFIX64 [RFC6052]. Organization specific values called Network-Specific Prefixes can also be used. Since multicast is potentially inter- domain, the use of well-known prefix for U_PREFIX64 is recommended. DHCP servers can also configure hosts with ASM_MPREFIX64, SSM_MPREFIX64 and U_PREFIX64 as in [I-D.ietf-softwire-multicast-prefix-option]. Note that U_PREFIX64 is also used in multicast data packet address translation. Source-specific multicast source address in multicast data packets coming from SSM sources MUST be translated using U_PREFIX64. 6.4. Supporting IPv4 Multicast at CE Router and lwB4 When MAP-E CE router is a NAT or NAPT box assigning private IPv4 addresses to the hosts, IP Multicast requirements for a Network Address Translator (NAT) and a Network Address Port Translator (NAPT) stated in [RFC5135] apply to IGMP messages and IPv4 multicast data packets. The same applies to lwB4s in lw4over6. On receiving multicast data packets, lwB4 or CE router MUST NOT modify destination IP address or destination port of the packets. Multicast UDP datagrams MUST be forwarded to the local LAN towards the host that is a member of this group. IGMP membership reports received at lwB4 or CE router may be sent upstream individually for any source multicast but for source specific multicast, e.g. IGMPv3, membership reports MUST be sent after IGMP aggregation. Sarikaya & Ji Expires April 2, 2015 [Page 19] Internet-Draft Multicast Support for MAP-E and Lw4over6 September 2014 7. Security Considerations 4rd Encapsulation Multicast control and data message security can be provided by the security architecture, mechanisms, and services described in [RFC4301], [RFC4302] and [RFC4303]. 4rd Translation Multicast control and data message security are as described in [RFC4607] for source specific multicast. 8. IANA Considerations TBD. 9. Acknowledgements TBD. 10. References 10.1. Normative References [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, March 1997. [RFC2629] Rose, M., "Writing I-Ds and RFCs using XML", RFC 2629, June 1999. [RFC1112] Deering, S., "Host extensions for IP multicasting", STD 5, RFC 1112, August 1989. [RFC2113] Katz, D., "IP Router Alert Option", RFC 2113, February 1997. [RFC2711] Partridge, C. and A. Jackson, "IPv6 Router Alert Option", RFC 2711, October 1999. [RFC3171] Albanna, Z., Almeroth, K., Meyer, D., and M. Schipper, "IANA Guidelines for IPv4 Multicast Address Assignments", RFC 3171, August 2001. [RFC4605] Fenner, B., He, H., Haberman, B., and H. Sandick, "Internet Group Management Protocol (IGMP) / Multicast Listener Discovery (MLD)-Based Multicast Forwarding ("IGMP/MLD Proxying")", RFC 4605, August 2006. [RFC4607] Holbrook, H. and B. Cain, "Source-Specific Multicast for IP", RFC 4607, August 2006. Sarikaya & Ji Expires April 2, 2015 [Page 20] Internet-Draft Multicast Support for MAP-E and Lw4over6 September 2014 [RFC3307] Haberman, B., "Allocation Guidelines for IPv6 Multicast Addresses", RFC 3307, August 2002. [RFC2491] Armitage, G., Schulter, P., Jork, M., and G. Harter, "IPv6 over Non-Broadcast Multiple Access (NBMA) networks", RFC 2491, January 1999. [RFC2473] Conta, A. and S. Deering, "Generic Packet Tunneling in IPv6 Specification", RFC 2473, December 1998. [RFC2765] Nordmark, E., "Stateless IP/ICMP Translation Algorithm (SIIT)", RFC 2765, February 2000. [RFC3022] Srisuresh, P. and K. Egevang, "Traditional IP Network Address Translator (Traditional NAT)", RFC 3022, January 2001. [RFC3810] Vida, R. and L. Costa, "Multicast Listener Discovery Version 2 (MLDv2) for IPv6", RFC 3810, June 2004. [RFC3376] Cain, B., Deering, S., Kouvelas, I., Fenner, B., and A. Thyagarajan, "Internet Group Management Protocol, Version 3", RFC 3376, October 2002. [RFC4213] Nordmark, E. and R. Gilligan, "Basic Transition Mechanisms for IPv6 Hosts and Routers", RFC 4213, October 2005. [RFC4301] Kent, S. and K. Seo, "Security Architecture for the Internet Protocol", RFC 4301, December 2005. [RFC4302] Kent, S., "IP Authentication Header", RFC 4302, December 2005. [RFC4303] Kent, S., "IP Encapsulating Security Payload (ESP)", RFC 4303, December 2005. [RFC5135] Wing, D. and T. Eckert, "IP Multicast Requirements for a Network Address Translator (NAT) and a Network Address Port Translator (NAPT)", BCP 135, RFC 5135, February 2008. [RFC6145] Li, X., Bao, C., and F. Baker, "IP/ICMP Translation Algorithm", RFC 6145, April 2011. [RFC6052] Bao, C., Huitema, C., Bagnulo, M., Boucadair, M., and X. Li, "IPv6 Addressing of IPv4/IPv6 Translators", RFC 6052, October 2010. Sarikaya & Ji Expires April 2, 2015 [Page 21] Internet-Draft Multicast Support for MAP-E and Lw4over6 September 2014 [RFC6224] Schmidt, T., Waehlisch, M., and S. Krishnan, "Base Deployment for Multicast Listener Support in Proxy Mobile IPv6 (PMIPv6) Domains", RFC 6224, April 2011. [RFC6333] Durand, A., Droms, R., Woodyatt, J., and Y. Lee, "Dual- Stack Lite Broadband Deployments Following IPv4 Exhaustion", RFC 6333, August 2011. [RFC6275] Perkins, C., Johnson, D., and J. Arkko, "Mobility Support in IPv6", RFC 6275, July 2011. [RFC7371] Boucadair, M. and S. Venaas, "Updates to the IPv6 Multicast Addressing Architecture", RFC 7371, September 2014. [I-D.ietf-softwire-map] Troan, O., Dec, W., Li, X., Bao, C., Matsushima, S., Murakami, T., and T. Taylor, "Mapping of Address and Port with Encapsulation (MAP)", draft-ietf-softwire-map-10 (work in progress), January 2014. [I-D.ietf-softwire-lw4over6] Cui, Y., Qiong, Q., Boucadair, M., Tsou, T., Lee, Y., and I. Farrer, "Lightweight 4over6: An Extension to the DS- Lite Architecture", draft-ietf-softwire-lw4over6-10 (work in progress), June 2014. [I-D.ietf-softwire-map-t] Li, X., Bao, C., Dec, W., Troan, O., Matsushima, S., and T. Murakami, "Mapping of Address and Port using Translation (MAP-T)", draft-ietf-softwire-map-t-05 (work in progress), February 2014. [I-D.ietf-softwire-4rd] Despres, R., Jiang, S., Penno, R., Lee, Y., Chen, G., and M. Chen, "IPv4 Residual Deployment via IPv6 - a Stateless Solution (4rd)", draft-ietf-softwire-4rd-08 (work in progress), April 2014. [I-D.ietf-mboned-64-multicast-address-format] Boucadair, M., Qin, J., Lee, Y., Venaas, S., Li, X., and M. Xu, "IPv6 Multicast Address With Embedded IPv4 Multicast Address", draft-ietf-mboned-64-multicast- address-format-06 (work in progress), September 2014. Sarikaya & Ji Expires April 2, 2015 [Page 22] Internet-Draft Multicast Support for MAP-E and Lw4over6 September 2014 [I-D.ietf-softwire-multicast-prefix-option] Boucadair, M., Qin, J., Tsou, T., and X. Deng, "DHCPv6 Option for IPv4-Embedded Multicast and Unicast IPv6 Prefixes", draft-ietf-softwire-multicast-prefix-option-07 (work in progress), September 2014. 10.2. Informative references [RFC3306] Haberman, B. and D. Thaler, "Unicast-Prefix-based IPv6 Multicast Addresses", RFC 3306, August 2002. [RFC3956] Savola, P. and B. Haberman, "Embedding the Rendezvous Point (RP) Address in an IPv6 Multicast Address", RFC 3956, November 2004. [RFC4291] Hinden, R. and S. Deering, "IP Version 6 Addressing Architecture", RFC 4291, February 2006. [I-D.ietf-softwire-dslite-multicast] Qin, J., Boucadair, M., Jacquenet, C., Lee, Y., and Q. Wang, "Delivery of IPv4 Multicast Services to IPv4 Clients over an IPv6 Multicast Network", draft-ietf-softwire- dslite-multicast-08 (work in progress), September 2014. [I-D.sarikaya-softwire-6man-raoptions] Sarikaya, B., "IPv6 RA Options for Translation Multicast Prefixes", draft-sarikaya-softwire-6man-raoptions-01 (work in progress), February 2013. [I-D.perreault-mboned-igmp-mld-translation] Perreault, S. and T. Tsou, "Internet Group Management Protocol (IGMP) / Multicast Listener Discovery (MLD)-Based Multicast Translation ("IGMP/MLD Translation")", draft- perreault-mboned-igmp-mld-translation-01 (work in progress), April 2012. Appendix A. Group Membership Message Translation Details IGMP Report messages (IGMP type number 0x12 and 0x16, in IGMPv1 and IGMPv2 and 0x22 in IGMPv3) are translated into MLD Report messages (MLDv1 ICMPv6 type number 0x83 and MLDv2 type number 0x8f). IGMP Query message (IGMP type number 0x11) is translated into MLD Query message (ICMPv6 type number 0x82) [I-D.perreault-mboned-igmp-mld-translation]. Destination address in ASM, i.e. IGMPv1, IGMPv2 and MLDv1, is the multicast group address so the destination address in IGMP message is Sarikaya & Ji Expires April 2, 2015 [Page 23] Internet-Draft Multicast Support for MAP-E and Lw4over6 September 2014 translated into the destination address in MLD message using [I-D.ietf-mboned-64-multicast-address-format]. Destination address in SSM, i.e. IGMPv3 and MLDv2 is translated as follows: it could be all nodes on link, which has the value of 224.0.0.1 (IGMPv3) and ff02::1 (MLDv2), all routers on link, which has the value of 224.0.0.2 (IGMPv3) and ff02::2 (MLDv2), all IGMP/ MLD-capable routers on link, which has the value of 224.0.0.22 (IGMPv3) and ff02::16 (MLDv2). Source address of MLD message that CE sends is set to link-local IPv6 address of CE's WAN side interface. Source address of MLD message that BR sends is set to link-local IPv6 address of BR's downstream interface. Multicast Address or Group Address field in IGMP message payloads is translated using [I-D.ietf-mboned-64-multicast-address-format] as described above into the corresponding field in MLD message. Source Address in IGMPv3 message payloads is translated using U_PREFIX64, the IPv6 unicast prefix to be used by SSM source. [RFC6052] defines in Section 2.3 the address translation algorithm of embedding an IPv4 source address and obtaining an IPv6 source address using a network specific prefix like U_PREFIX64. At the BR on its upstream interface or at the CE on its LAN interface, IPv4 addresses are extracted from the IPv4-embedded IPv6 addresses. Maximum Response Time (MRT) field in IGMPv2 and IGMPv3 queries are translated into Maximum Response Delay (MRD) in MLDv1 and MLDv2 queries, respectively. In the corresponding MLD message, MRD is set to 100 times the value of MRT. At the BR on its upstream interface or at the CE on its LAN interface, MRT value is obtained by dividing MRD into 100 and rounding it to the nearest integer. IGMP messages are sent with a Router Alert IPv4 option [RFC2113]. The translated MLD message are sent with a Router Alert option in a Hop-By-Hop IPv6 extension header [RFC2711]. In both cases, 2-octet value is set to 0. Authors' Addresses Behcet Sarikaya Huawei USA 5340 Legacy Dr. Building 175 Plano, TX 75024 Email: sarikaya@ieee.org Sarikaya & Ji Expires April 2, 2015 [Page 24] Internet-Draft Multicast Support for MAP-E and Lw4over6 September 2014 Hui Ji China Telecom NO19.North Street Beijing, Chaoyangmen,Dongcheng District P.R. China Email: jihui@chinatelecom.com.cn Sarikaya & Ji Expires April 2, 2015 [Page 25]