The BGP/MPLS VPN solution is built on the peer model described in the previous section. This should not come as a surprise, for two reasons. First, we have seen that PE-based VPNs have attractive properties such as simple routing from the customer's point of view and easy addition of new VPN sites. Second, we have seen that early PE-based solutions were limited by the fact that traffic traveled as IP in the core. Tunneling would eliminate this limitation, and MPLS can provide the necessary tunnels.

The BGP/MPLS VPN model was first published in informational RFC 2547 [RFC2547], documenting a solution developed at Cisco. Following the success of 2547 VPNs there was a desire from some service providers to make it into an IETF standard. A new working group was started in the IETF, called ppvpn (for provider-provisioned VPNs). One of the work items of the group was to standardize MPLS VPNs, and the internet draft that resulted from this work was named 2547bis. In the industry today, BGP MPLS/VPNs are often called 2547bis for this reason although in the meantime it has been standardized as RFC 4364 [RFC4364]. The ppvpn Working Group undertook work in both the L2 and L3 spaces, and was later split into the l2vpn and the l3vpn Working Groups [L3VPN, L2VPN].

In the following sections, we will build the BGP/MPLS VPN solution step by step, hopefully shedding light on some of the design decisions taken. Before we can start, let us remember the goals ...

Get MPLS-Enabled Applications: Emerging Developments and New Technologies now with the O’Reilly learning platform.

O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.