[e2e] e2e principle..where??....

[Manish Karir]

[please limit future discussions to e2e list only...I accidentally posted
 to the tsvwg as well..I apologize for the cross posting...:(  ]

If I read your comments correctly, then one would infer that a proxy 
written at the appliction level(say a web proxy) does not violate 
the e2e principles??... the client---proxy and proxy---server connections
are the connections that "directly exchange" data....and for those single 
links the e2e principles hold.  After all quiet a few caches are infact
built as "caching proxy systems"(squid being a good example).  So then
should we also consider 
appliction level proxies as systems that do not break e2e principles..???
(so long as the proxies are being faithful and not changing
data/headers??)

manish karir 

On Fri, 1 Jun 2001, Eric A. Hall wrote:

> But are the intermediaries being faithful? or are they changing headers
> and doing other things to the data?
> 
> The true end-points are the nodes that exchange data directly, not the
> ultimate source or destination for the data.
> 
> We can say that there are tremendous benefits to caches to have them
> intercept HTTP data, but that is a convenience, not a requirement of the
> client-cache connection. The cached data could just as easiily have been
> populated by a filesystem or a database query as it was from HTTP. The
> ultimate end-to-end link will be between the client and the cache, not the
> client and the server, and so the protocol semantics and other aspects of
> the cache-server link are irrelevant.