RE: Models & Scenarios

[DonE@activate.net]

> -----Original Message-----
> From: Phil Rzewski [mailto:philr@inktomi.com]
> Sent: Monday, March 05, 2001 2:35 PM
> To: cdn@ops.ietf.org
> Subject: Models & Scenarios
> 
> 
> Hi team. I just wanted to send along this companion message a 
> background 
> the changes I've made to the Models & Scenarios documents.
> 
> The overall goal is to "float a balloon" of some high-level 
> ideas we've 
> seemed to have rough consensus on since the last BOF meeting. 
> The first of 
> these was trying to generalize from "CDNs" and "CDN Peering" 
> to "Content 
> Networks" and "Content Internetworking". Other than just a change in 
> wording, this also implies some change in scope.
> 
> For example, in the Models document, I added some sections showing 
> historical types of "Content Networks", such as hierarchical 
> caching and 
> server farms. In the Scenarios document, this meant naming 
> some specific 
> traits/refinements of certain peer-worthy Content Networks, 
> some of which 
> I've given names to like "Access Content Networks", 
> "Publishing Content 
> Networks", and "Brokering Content Networks". While I got some 
> preliminary 
> support from some folks on this approach, I'm curious to see how it's 
> received by the wider audience. I have historically seen two 
> approaches: 1) 
> Think of everything as a CDN, think of refined cases as 
> degenerate cases of 
> CDNs, and 2) Cite existing examples that don't look much like 
> what's called 
> a CDN today and call them out by name. So this is a stab at 
> the latter.

My interpretation of section 2 was not quite as strong as you state in
item 2) above, but it puts CNs in broad categories (as you state "networks
gravitate toward particular subsets of the Content Internetworking
interfaces".  The ultimate goal is to understand the concept of CNs and
enumerate all possible scenarios (as done in section 3), so that the
requirements documents define a common protocol that can be used by all
types of CNs.

Don Estberg