Workflows Needed To Create an IP VPN
From the point of view of the sites in each VPN, the CE-PE links in Figure 1 are the key elements. These links carry IP traffic into and out of a given VPN. If we assume that this link is configured with the appropriate technologies, on the PE interface side, as we saw above, the network manager must do the following:
Configure the virtual routing and forwarding instance (VRF).
Configure the route distinguisher (RD).
Configure the route target (RT).
Once this data is configured on all associated PEs, the VPN is essentially complete.
Additional parameters on the CE-PE link may need to be configuredDiffServ, policing, etc. But, for our purposes, the user-facing setup consists of configuring just the adjacent interfaces on the CE and PE routers.
From the service provider's perspective, the core must also be configured. This process involves setting up routing and signaling protocols as illustrated in the center of Figure 1. One additional core task is creating the PE-to-PE LSPs (LSP A and LSP B in Figure 1). These LSPs transport the MPLS-encapsulated IP traffic. Once the LSPs have been created (either manually, using RSVP-TE, or automatically, using LDP) and enabled, traffic can then flow between the VPN sites.
So, in broad terms, we can say that the workflows required to create an IP VPN consist of steps 1 to 3 above as well as the creation of one or more traffic-engineered tunnels or LSPs; that is:
Configure the PE node interfaces.
Create the core LSPs.
Tie the VRFs to specific LSPs.
Clearly, there are other setup tasks, such as configuring the core nodes and interfaces. The emphasis here is on the workflows needed to bring up the end-user service.
Given these simple workflows, let's take a look at how the existing IP VPN MIBs stack up. Do the MIBs facilitate the workflows?