Re: Latest NHRP draft

[Bruce Cole <[e-mail]>]

> Can you elaborate on how rate limiting would prevent
> multiple VCs from occuring when the NHRP request goes through multiple
> NHSs towards its destination.

First: I was not arguing that rate limiting prevents multiple VCs in general.

In this particular case, multiple VCs can be avoided through rate limiting.
Specifically: router B receives an NHRP request from router A.  Depending
upon packet ordering, B may or may not have already received the original
IP packet which A is trying to forward to D.  If the original IP packet
has not yet been received when B receives the NHRP request, then B
may suppress sending an NHRP request of its own, to provide rate limiting.  If
the original IP packet came first, then B probably already sent an NHRP
request of its own.  When the NHRP packet from A comes along, B may then drop
A's request due to rate limiting.

C behaves similarly.  End result is only 1 NHRP request to resolve, so only
1 VC is established.

> Does this method guarantee that multiple VCs will not occur?

No.

> If the source has multiple NHRP requests within a short period of time,
> would rate limiting affect the subsequent requests?

Yes or no depending upon rate limiting policy.  If the policy is to
allow X NHRP packet within interval Y, then subsequent requests may or
may not be dropped depending upon the values for X and Y.