[Robustness] Modifying Annex E to support Robustness?

Qiaobing Xie xieqb at CIG.MOT.COM
Mon Apr 10 17:43:12 EDT 2000


"Paul E. Jones" wrote:
...
> I am definitely giving consideration to using a modified Annex E/H.323.
>
> Taking a broader view, I think that a connectionless approach is the right
> solution.  I am considering both how Annex E or SCTP might satisfy the
> requirements .

Similat to TCP or SCTP, Annex E achieves its data reliability by means
of sequence numbering and retransmission of lost messages. This requires
both sides initialize sequencing and retransmission state and keep
tracking those state during data transfer. From this view, Annex E can
hardly be categorized as connectionless when used for reliable data
transport.

I guess you were probably more referring to the "fast set-up" feature of
Annex E, ie., you can send a user message to a new destination without
explicitly going through a set-up procedure (while the sender and
receiver stack will still internally go through set-up to allocate
resource for the state information of this new connection and start
tracking the sequence number, etc). This can be done with SCTP by
piggy-backing the first user message with the INIT and INIT-ACK. The
problem with this approach (the same is true to Annex E) is that you
will lose most of the security protection on the destination end. People
may argue that you can always run Annex E over IPSEC to get the
protection, but IPSEC session is connection-oriented; before you can
deliver the first user message, the IPSEC itself will need to establish
an security association between the two endpoints.

-Qiaobing



More information about the sg16-avd mailing list