The Routing Over Large Clouds Mailing List Archive by date

[yakov@watson.ibm.com]

Ref:  Your note of 3 Mar 95 15:55:00 -0500


John,

>Moreover, NHRP Fabric Mode procedures require that the next-hop
>determined by routing be resolved by some means other than NHRP.

That brings a bigger issue -- where does an NHS gets its information
from ? If the information is static, then the use of NHRP in Fabric
Mode amounts to nothing short of static routing. If only part of a path
from a source to a destination is constructed using this information
then it is relatively easy to construct examples that show how a mix of
static routing (resulted from using NHSs that rely on static
information) and dynamic routing (for the parts of the path that is
outside of the ATM fabric) produces persistent forwarding loops.

>why must we continue to cling to NHRP as the only possible solution ?

I certainly hope we wouldn't. NHRP may be viewed as "a solution" that
doesn't cover all the cases (specifically the router-to-router case).
Other solutions, especially for the router-to-router case, must be
possible.  In fact, two recent I-Ds by Curtis, Kannan and myself
provide one possible solution for certain scenarios of the
router-to-router case. So, I would see little or no reason for a router
that participates in exterior routing to do NHRP as well.

Yakov.