W3C home > Mailing lists > Public > ietf-http-wg@w3.org > April to June 2007

Re: 305 Use Proxy deprecated?

From: Adrien de Croy <adrien@qbik.com>
Date: Sat, 16 Jun 2007 22:19:25 +1200
Message-ID: <4673B92D.5060508@qbik.com>
To: Henrik Nordstrom <henrik@henriknordstrom.net>
CC: HTTP Working Group <ietf-http-wg@w3.org>


It's certainly interesting.

Reading through the posts about it from 1997, it originally had quite a 
different perceived use I think, since it was foreseen that it would be 
used by proxies more than by origin servers I think... at some stage use 
by proxies was banned.

I can see a couple of useful scenarios where it would be useful for a 
server to tell a UA to use a proxy, but most of these use cases would be 
where it was a proxy telling a client.

for instance a distributed cache, where one proxy in a cluster which 
shares a common cache index gets a request which is hosted on another 
proxy in the cluster - to be able to redirect the UA to the proxy that 
holds the file would be one use case (rather than the proxy simply 
diverting through to the other cluster member).

The other case could be load balancing.

But anyway, not a biggie.  I think given the extreme lack of wording on 
the code in the spec, perhaps some words could be added to discourage 
anyone from wasting their time implementing it in a server.

Regards

Adrien


Henrik Nordstrom wrote:
> fre 2007-06-15 klockan 17:16 +1200 skrev Adrien de Croy:
>
>   
>> can't find any browser that supports this.
>>     
>
> I fail to find use cases where 305 makes much sense, and several
> security concerns.
>
> proxies is normally in the users administrative realm, while 305 is only
> issued by origin servers (MUST NOT be issued by proxies..). This has
> notable implications on authentication, making 305 a very odd case.
>
> To me it simply doesn't make any practical sense. The normal redirects
> suits those use cases where a server wants the request to be sent via
> another server much better I think.
>
>   
>> So looks like the main browsers (haven't tried Safari) have de facto 
>> deprecated it.
>>     
>
> Yes. Or rather all browsers except Opera choose to ignore 305
> completely.
>
> Also I don't think there is many proxies implementing 305.
>
> For 305 to work out the way it is designed it must be handled as a
> hop-by-op operation, redirecting the last hop via the indicated proxy.
> It should not go up to the user-agent if the user-agent is already using
> a proxy.
>
> Regards
> Henrik
>   
Received on Saturday, 16 June 2007 10:19:30 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 27 April 2012 06:50:10 GMT