RE: Executive summary of: RE: hard questions: request routing

["Cain, Brad" <bcain@cereva.com>]

> -----Original Message-----
> From: Dmitri Krioukov [mailto:dima@krioukov.net]
> Sent: Wednesday, April 04, 2001 12:29 AM
> To: Oliver Spatscheck
> Cc: cdn@ops.ietf.org
> Subject: RE: Executive summary of: RE: hard questions: request routing
> 
> 
> Has the IETF started voting? -- 1 :)
> 
> I missed the point when the metric was dismissed, though :(

in both the mailing list and the request routing requirements
we dismissed a single metric

> Another thing is what work is being done -- on requirements
> (loop free) or on methods/protocols (options below)? Anyway,
> restriction of topology goes well beyond both up to the
> inter-CDN architecture.

you have missed the point... we are not restricting TOPOLOGY,
but FORWARDING mechanism... think the difference between
"hop-by-hop" routing vs. source-routing...  source-routing
works quite nicely for CDNs especially DNS based ones... it
also allows each authoritative CDN to make its own decision

if people want "hop-by-hop" routing then we may have to
restrict the topology (unless we want to go down the
rathole of diffusion updates)... i'd argue very strongly
that more than two hops of request routing breaks the
whole CDN value proposition... do you have an argument
otherwise?

-brad