W3C home > Mailing lists > Public > ietf-http-wg-old@w3.org > May to August 1997

Re: HTTP/1.1 -- who disconnects the session?

From: David W. Morris <dwm@xpasc.com>
Date: Thu, 5 Jun 1997 21:46:57 -0700 (PDT)
To: Josh Cohen <josh@netscape.com>
Cc: Michael Giroux <mgiroux@worldnet.att.net>, "Http-Wg (E-mail)" <http-wg@cuckoo.hpl.hp.com>
Message-Id: <Pine.GSO.3.96.970605214128.9938A-100000@shell1.aimnet.com>
X-Mailing-List: <http-wg@cuckoo.hpl.hp.com> archive/latest/3417

On Thu, 5 Jun 1997, Josh Cohen wrote:

> I beleive the connection is dropped after the client's
> final request.  
> The final request is indicated by passing 
> the connection: close header with its final request..
> (feel free to correct me if I'm wrong. )

Last I checked, either side closes the connection when it chooses. 
Choice from a server perspective is expected to relate to available 
resources.  Choice on the client would be expected to relate to when the
client heuristically decides a connection isn't needed.

Either side can use the connection close req/resp to gracefully terminate
the connection. A decision that this point has arrived is still necessary.

There was a recent paper reporting on some performance studies made by the
folks at W3C. There has also been a recent suggestion regarding close wait
issues that suggested a preference for which side should close the
connection first.  Sorry, but I forget, check the archives for this list.

Dave Morris
Received on Thursday, 5 June 1997 21:52:27 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 14:40:20 UTC