W3C home > Mailing lists > Public > ietf-http-wg@w3.org > July to September 2006

RE: HTTP/1.1 pconns to 1.0 servers

From: Paul Leach <paulle@windows.microsoft.com>
Date: Fri, 22 Sep 2006 22:36:20 +0000
Message-ID: <76323E9F0A911944A4E9225FACFC55BA02257C36@WIN-MSG-20.wingroup.windeploy.ntdev.microsoft.com>
To: Mark Nottingham <mnot@yahoo-inc.com>
CC: <ietf-http-wg@w3.org>, Jeffrey Mogul <Jeff.Mogul@hp.com>




In HTTP/1.0, persistent connection was not the default, and didn't even
need to be supported.

Hence, unless an HTTP/1.1 client somehow knew a priori that it was
speaking to an HTTP/1.1 server, I don't see any way to escape the need
for HTTP/1.1 clients to send "Connection: keep-alive" if it wants to
support persistent connections whenever possible. And they will have to
keep doing so until it makes sense for them to believe that support for
HTTP/1.0 servers can be dropped.

(Well, almost. An HTTP/1.1 client doesn't have to send Connection header
-- if it doesn't, it just needs to be prepared for the server to
indicate the end of the entity-body by dropping the connection. However,
this would only be good engineering when the number of HTTP/1.0 servers
drops to a small enough level that not using persistent connections has
no significant impact on Internet load.)

-----Original Message-----
From: ietf-http-wg-request@w3.org [mailto:ietf-http-wg-request@w3.org]
On Behalf Of Jeffrey Mogul
Sent: Friday, September 22, 2006 2:15 PM
To: Mark Nottingham
Cc: ietf-http-wg@w3.org
Subject: Re: HTTP/1.1 pconns to 1.0 servers


    What was the intent here? I can't imagine that it was the intent of

    the WG to require HTTP/1.1 clients to send Connection tokens in  
    perpetuity for backwards compatibility.
    
Received on Saturday, 23 September 2006 21:03:47 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 27 April 2012 06:49:46 GMT