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

[David P. Reed]

I stand corrected by Ran Atkinson about @Home using transparent 
caches.  Not being an @Home user, I relied on statements made to me by 
others, which may well be incorrect.

To clarify and extend my point, Ran, which was indeed intended to be purely 
technical except for the parenthetical remark about @Home, the "proxy" 
approach seems to me also to be more-or-less compliant with the end-to-end 
argument for the same reason that Akamai is, except perhaps when it is 
"mandated" as a condition of access.

But this is only true because correctly behaving "proxies" have been 
accepted as part of the WWW architecture and protocol set.  (that the 
mechanisms don't work well and are not properly implemented in many cases 
should probably be viewed as a pervasive set of bugs and design flaws in 
the WWW protocol set, browsers, and servers).  WWW server code must be 
designed so that it works properly with proxies of this sort (because 
logically the proxy is part of the client functionality according to the 
WWW specs).

- David
--------------------------------------------
WWW Page: http://www.reed.com/dpr.html