The MPLS WG Archive

["David Allan" <dallan@nortelnetworks.com>]

Title: RE: [Fwd: I-D ACTION:draft-pan-rsvp-te-restart-01.txt]

Ping:

Comments in line....

<snip>
        > You (and others, my
        > point is not specific to this I-D) are eliminating all forms of fate
        > sharing between the control and data planes, despite (in the MPLS case
        > vs. GMPLS) the use of common resources in the control and data plane
        > adjacencies.

        David,

        What has been proposed in all forms of graceful restart is the message
        format on the wire. How you implement the routers to support these
        messages is totally up to the developers.

Seems to me that interoperability and overall network survivability would mandate at least some agreement on the semantics associated with this stuff....

        > This may suggest that control plane failure is today significantly more
        > frequent than data plane or link failure and there is a real problem to
        > solve, but adding restart mechanisms to everything in the control plane
        > is not the complete answer.

        You are right. There are several solutions to help routers/networks in
        case of failure condition. Graceful restart is to solve the problem when
        control plane dies, but data forwarding is up. Fast reroute is to solve
        the problem when data forwarding is in trouble. LSP-ping (please don't
        start all those architectural beatification argument again) is to detect
        data plane black-hole problem....

Secular issues aside ;-), what you describe above is a pragmatic response to improving the reliability of a particular application of MPLS, and that's OK, but I trust you aren't suggesting that the above is a complete solution....

cheers
Dave