Cell Relay Archive

[Yu Ning <yuning@mindless.com>]

Hi friends,

In my previous posting about MPOA's scalibility question, Bert and other
friends' answer mainly focus on the NHRP part of MPOA, they think that:
positioned at the edge of the whole ATM cloud, NHS will maintain a large,
and non-hierarchical IP-ATM mapping table, and this is the key defect to
scalibility. What's more, friends have point out that the solution to MPOA
scalibility problem is I-PNNI, and MPLS.

After a week study of I-PNNI(together with PAR), I found that, in the
I-PNNI model, NHRP is still needed for address resolution, because
I-PNNI advertise reachibility information of IP and ATM seperately. Only
in certain case, IP reachibility info will contain IP-ATM pair (i.e. in
care of reachibility), so an address resolution mechanism is still
needed in I-PNNI. Then, it seems I-PNNI hasn't solve the problem of
NHRP. Am I right? :-)?

What's I-PNNI solved, it seems, is the scalibility of ATM cloud. Because
in MPOA Spec, the ATM network signalling/routing part is based on
UNI 3.1/4.0, which in fact utilize no dynamic routing protocol. While
in I-PNNI, nodes use PNNI routing protocol to exchange routing info,
and this lead to the scalibilty of underlying ATM network. 

Then I plan to study MPLS for weeks, and try to find out whether it
solve the problem mentioned by friends' answer.

BTW, 
It's seem that it's quite cold in the research of I-PNNI, and all the
resource I get about it is written in 1996(i.e. ATM-F contributions).
And It seems the main stream of I-PNNI study have move to MPLS?
right?(just from a brief view of writters of two group :-)).

Thanks for any comment.

Yu Ning.

------------------------------------------------------------------------------*
Yu Ning
ATM R&D Center of 
BUPT (Beijing U. of Posts&Telecom)
Beijing  (ZIP: 100876,  MBox: 147#),   P. R. China
-------------------------------------------------------------------------------