Juniper JUNOSE, Configuration Manual

Page: 428 / 738

preferred to option A. For inter-AS options B and C, you must explicitly configure
MPLS on all the inter-AS links.
Inter-AS Option A
Figure 77 on page 392 illustrates the first method, where you create a VRF for each
VPN on each AS boundary router.
Figure 77: Inter-AS Topology with VRFs on Each AS Boundary Router
Within each AS, routes are announced by internal MP-BGP and the data packets are
forwarded across an MPLS tunnel. You create a logical connection such as an ATM
VC between each pair of VRFs (on separate AS boundary routers); these logical
connections can share the same physical connection. The following factors limit the
scalability of this method:
■
All inter-AS VPN routes (potentially a very large number) must be stored in the
BGP RIBs and IP routing tables on the AS boundary routers.
■
You must configure VRFs on each AS boundary router.
MPLS tunnels are unidirectional; Figure 77 on page 392 shows only the tunnels
established to carry traffic from ASBR 2 to PE 1 and from PE 4 to ASBR 3. Note that
ASBR 2 and ASBR 3 are both also PE routers. In that sense, ASBR 2 treats ASBR 3
as a CE router, and ASBR 3 treats ASBR 2 as a CE router.
Inter-AS Option B
The second method is known as inter-AS option B or 2547bis option B, after IETF
RFC 4364—BGP/MPLS IP Virtual Private Networks (VPNs) (February 2006). This
method uses BGP to signal VPN labels between the AS boundary routers (Figure 78
on page 393). 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 provides greater
scalability, because only the BGP RIBs store all the inter-AS VPN routes.
392 ■
Providing IPv4 VPN Services Across Multiple Autonomous Systems
JUNOSe 11.0.x BGP and MPLS Configuration Guide