The JUNOSe software supports inter-AS services for IPv6 VPNs in addition to IPv4 VPNs. See Providing IPv4 VPN Services Across Multiple Autonomous Systems for more information about inter-AS services and IPv4 VPNs.
The JUNOSe software currently supports only 2547bis option B for IPv6 VPNs. This method—described in RFC 4364—BGP/MPLS IP Virtual Private Networks (VPNs) (February 2006)—uses BGP to signal VPN labels between the AS boundary routers (Figure 81). The base MPLS tunnels are local to each AS. Stacked tunnels run from end to end between PE routers on the different ASs. This method enhances scalability, because only the BGP RIBs store all the inter-AS VPN routes.
Figure 81: Inter-AS IPv6 VPN Services
In Figure 81, the base tunnels between the PE routers are established in the IPv4 core networks with LDP or RSVP. The PE routers advertise IPv6 prefixes from the CE devices within their respective ASs as VPNv6 prefixes with MP-IBGP. For example, PE 1 advertises the CE 1 prefix 6001:0430::/48 over to PE 2 in its MP_REACH_NLRI attribute. The next-hop attribute in the update message is the PE 1 loopback address—the IPv4-mapped IPv6 address, FFFF::220.127.116.11/128.
PE 2 advertises 6001:0430::/48 by means of MP-EBGP to PE 3. The prefix is sent as a VPNv6-labeled prefix (2002:0202/48), with the default BGP next hop being the IPv4-mapped IPv6 address of the IPv4 interface going to PE 3.
For inter-AS services, in contrast to intra-AS services, JUNOSe software supports both IPv4 backbone and IPv6 backbone types of BGP next-hop encodings. The default BGP next-hop encoding used for IPv6 VPN inter-AS services is the one specified for the IPv4 backbone where IPv4-mapped IPv6 addresses are used. Alternatively, you might also configure the IPv6 backbone type of BGP next-hop encoding by configuring route maps that use native IPv6 addresses for the BGP next hop.