Inverse ARP over Unidirectional Virtual Circuits
draft-heinanen-inarp-uni-01

Note: alex posted comments/questions on this ID  back in June, 2002.

http://cell-relay.indiana.edu/mhonarc/mpls/2002-Jun/msg00023.html
Juha:

Please find below some comments/questions I have on this draft:

1. The document seems to assume the NBMA model, i.e.,
  a set of LSPs is abstracted as a LIS, independent from
  the IP/MPLS domain which LSPs are established through.
  I think this should be spelled out.
 
2. How are the ARP messages encapsulated in case of MPLS LSPs,
  especially considering their (LSPs') uni-protocol nature?

3. What value is used for the Hardware type field in
  case of MPLS?

4. How are the *ha fields encoded? Any implications for
  the case of interface-specific label space?

5. What happens if a request or a reply gets dropped?
  How often should the request be retransmitted?

6. The scalability section needs more work, I think.

  It is not enough to say that transmission of the request should
  be randomly delayed unless you specify the time range. Otherwise
  we can have a situation where a node is brought up or gets connected
  to the cloud and every remote node sends exactly one request
  within a short period of time, but due to the total number of
  remote nodes we still have O(n^2) messages (requests +
  replies)? You also need to make sure that even in the situation
  where the requesters do not behave properly (or are malice) and
  you receive a lot of requests back to back, reply generation is
  still controlled...

--
Alex

State Changes to New Version Needed (WG/Author)                    from Reading List                                      by scoya

State Changes to Reading List                                      from AD to write --Don't publish                      by scoya