IS-IS route leaking question

[4everlearning]

Hi Guys,

Can somebody explain to me what is the following statement mean (which is a part of Cisco IS-IS documentation):

"In MPLS VPN environments, reachability information for the loopback addresses of every Provider Edge (PE) router is needed. Leaking routes for the PE loopback addresses allows a multiarea hierarchy to be used in MPLS VPN implementations".

The statement is not clear, why is that a must in MPLS VPN environment? 

Thanks,

[rais · Accepted solution]

By default, L2 routes are not distributed into L1. Instead, closest L1/L2 can receive all L1 traffic destined for any destination not within this specific L1 area.  

For MPLS VPNs, all PEs in the network need to be able to reach each other. This can be done:

1. by placing all PEs in the same L1 or L2 area. [no hierarchy]
2. by leaking PE routes from L2 area into L1 areas. [multi-area hierarchy]

In MPLS, LSPs better know all specific routes to reach the destination PE. This gives you some form of traffic-engineering to utilize multiple/optimal paths vs. single path to exit out.

HTH.

[4everlearning]

Thanks Rais, that makes sense! 

Just a small clarification on point number 2. Can multi-area hierarchy be achieved without leaking by making sure that all PEs are configured as L!/L2 (Not L1) ? since L1/L2 routers will be able to communicate with each other even though they are part of different areas. or even with this scenario, leaking is required? 

Regards,

[rais]

That's default for Cisco routers. L1/L2 will have complete knowledge of the network already. No leaking required.

Thanks.