Comments on draft-ietf-v11-spec-00

Hi, first let me introduce myself. I am Peter Churchyard and I work
in the firewalls group for Trusted Information Systems. I wrote the http-gw
that is in the firewall toolkit and our Gauntlet (TM) Internet firewall.

As an implementer of http proxies, I have alot of interest in getting
HTTP 1.1 to provide implementable and correct mechanisms.

I have a number of comments and after a quick scan of the archive at
www.ics.uci.edu have not seen any major discussion of.

Section 8.4 POST (and others)

This section discusses a two phase process for the processing of POST requests.
Part of the process is a five second pause.

Use of this type of mechanism cannot work in general! It is not possible
to design a reliable mechanism where the lask of a response in a timeframe
such as this is part of the protocol.

Due to network congestion, differing link speeds, network buffering etc
The request can be 'seen as sent' by the client but will not have arrived
at the origin server.

The only way to deal with this is to require a response (in reply to a
header in the request) if the client does not get the response in a suitable
timeframe it may decide that the server does not understand the request and
can safely drop the connection and try to reformulate the request in a way
that may be suitable.

On the other hand, the entire reason for this functionallity may be broke, and
the feature is not required at all.

Pete.
P.S. I would be happy to contribute my expertise to this project..

-- 
The TIS Network Security Products Group has moved!
voice: 301-527-9500 x123 fax: 301-527-0482
2277 Research Boulevard, 5th Floor, Rockville, MD 20850

Received on Sunday, 21 January 1996 12:38:20 UTC