[Tsvwg] Re: [e2e] e2e principle..where??....

[RJ Atkinson]

At 11:09 02/06/01, David P. Reed wrote:
>The "transparent cache" approach (pioneered by @Home) where the network itself spoofs HTTP and tries to second-guess what the user wants - that is not end-to-end, and creates huge secondary problems (since the server and client don't know that caching is going on, they may assume the client is getting up-to-date information, whereas the out-of-date cached information is being fed).

CORRECTION:
        @Home does not now have a "transparent cache" and 
never did have a "transparent cache".  @Home has now and 
has had configured web proxys all over its network, in many
headends and nearly all Regional Data Centres.

        However, @Home's web caches all rely on the user's instance 
of Netscape or IE being configured to use the web proxys.  As an 
@Home user for several years, I never used their caches.  A 
non-trivial number of @Home users did not use the caches as 
of the last measurements that I saw.  At @Home installation time, 
most user's web browsers were usually (not always) reconfigured 
to use the proxys by the installer.  Nothing prevents the user 
from using EDIT/PREFERENCES to change those settings to whatever 
the user wants.

        Mr. Reed's incorrect speculation about @Home's practices 
isn't advancing the technical conversation here, IMHO, but the
record did need to be straightened out on these lists.

Ran
rja at inet.org