W3C home > Mailing lists > Public > uri@w3.org > May 2008

Re: Different links, with the same cached page.

From: Al Gilman <Alfred.S.Gilman@IEEE.org>
Date: Sat, 3 May 2008 15:42:08 -0400
Message-Id: <089214B1-E068-44FF-AF87-9E03F92324F9@IEEE.org>
Cc: uri@w3.org
To: "F." <dev001@pas-world.com>


On 3 May 2008, at 8:45 AM, F. wrote to <www-html@w3.org>:

>
> Hello,
> I am trying to use different hosts to return a resource.
> I select on the fly the URI of the resource between different  
> available
> host at this moment (or random host).
> But I see a problem, do not see a very good choice, if I am using HTTP
> cache, because page is different when URI of resource change, page  
> never
> should be fresh.
>
> This problem could be solved with java-script but is not robot
> compliant.
> I think that in HTML do not exist vars, enumerations, I think that is
> not posible, ... :)
> Anyone know method, "spiders compliant", that I could use to make  
> use of
> different host URI with same public cached page.

The content distribution problem you are considering is the basis of
a rather well developed and stable market for out-sourced  
distribution services.

Current practice is either to expose the mirror choice to the user  
for their
manual selection (common in software distribution) or to deal with the
spatial options automatically between the HTTP layer and DNS (common in
media distribution).  Either way requires a network of geographically
distributed cooperating mirror sites.

The basic outlines of one solution are discussed in the WikiPedia page
on Akamai

http://en.wikipedia.org/wiki/Akamai_Technologies

Al

>
>
> -- 
>
>
>
>
> --
> Publicidad HTTP://www.pas-world.com
>
>
Received on Saturday, 3 May 2008 19:42:48 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 13 January 2011 12:15:41 GMT