RE: draft-lee-mpls-path-request-01

["Naidu, Venkata" <Venkata.Naidu@Marconi.com>]

Hass:

-> So it would query X and Y's Path Computation Entities
-> to get the routes across those areas. 

 Then effectively, an internal router getting all the 
 knowledge about external areas/routes. This approach 
 will have scalability problem.

-> But as the draft now stands,
-> A has to specify a single ingress to and egress from each area.

 Yes! but the presented approach still achieves link 
 disjointness if not nodedisjointnes. (link failures are 
 more frequent than complete node/ABR failures) Isn't it?

--Venkata Naidu