ietf-openproxy
[Top] [All Lists]

RE: draft-tomlinson-opes-model-00.txt

2001-07-27 11:27:57
--snip--snip--


yes, i agree with reinaldo,  OPES exists independent of CDI


HERE I MADE ADD,

The concept of edge services as an overlay is ok, however, 
the way that is presented in the draft is too restrictive for 
the following reason(s)

Agreed. I think your wording is better. My whole point is that overlay
should be better defined/expanded, not removed from the doc.

 - it assumes that each overlay is independent with a 
seperate authorative entity.
 ------ here, we can extend the concept of an edge services 
overlay to broder range, there is no restriction for several 
access providers or ISPs to cooporate such that their edge 
services become an extended overlay (an overlay of overlays).

 - the overlays can also be combined, basically, a content 
overlay can be defined that provide content to one or more 
edge services overlays (or from above a virtual edge services 
overlay that is composed of various edge services overlays).

 - the same can be said about client overlays. (here, an 
overlay can be defined that allow clients to roam across 
various edge services overlay(s)).

in fact, the edge services overlay model can be used as an 
alternative to the cdi peering model,
in this case, the content provider(s) can be though off as 
having their own overlay(s). the edge services as another 
overlay and the clients as an overlay.

Very Good Point.
 

for an OPES intermeiary (surrogate) to act on behalf of a 
content source, it has to subscribe to the content overlay 
(similar analogy can be used about surrogates or caches).


we can discuss the analogy further if there is interest

hope the above helps


abbie








<Prev in Thread] Current Thread [Next in Thread>