ISIS MPLS PE reachability (route leaking)

[Michael Bartholomæussen]

I understand that PE routers must have reachability in a MPLS VPN environment, but when using ISIS isn't the default route back into a L2 area sufficient? is this mainly to avoid sub-optimal routing or is there any other reason for leaking PE routers into L1 areas?

Regards,

Michael 

[Julio E. Moisa]

Hi Michael,

ISIS like OSPF has a type of backbone and this backbone works over Level-2 infrastructure, now when a L1 Router is connected to this backbone, it enables the ATT bit and default route is advertised from the L1 edge router to other L1 routers behind. It is useful to enable the communication between L1 segments.

If you are using MPLS over ISIS basically it will be configured on the ISP side. Rarely ISIS is seen configured on the client side. 

Now you can leak L1 prefixes over L2 leves and vice versa using leak maps. 

[Michael Bartholomæussen]

Hi Julio,

I have read, that with route leaking in ISIS it enables MPLS PE router reachability. If i refer to this post https://supportforums.cisco.com/t5/wan-routing-and-switching/mpls-vpn-isis-loopback-leaking/m-p/2519716#M238708

the main question is kinda the same, but the answer in the post seems a bit unclear. Perhaps you can elaborate it a bit further?

If we think of a topology like this,(PE1-L1) - L2 - (L1-PE2). Assuming that there is a MPLS VPN between PE1 and PE2, wouldn't default routing via the ATT bit, be sufficient to achiveve reachability between the PE routers. The route would be subjectable to sub-optimal routing, but it would work.