[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: new to cdn internetworking



Title: RE: new to cdn internetworking


> -----Original Message-----
> From: Christian Kuhtz [mailto:ck@arch.bellsouth.net]
> Sent: Tuesday, November 13, 2001 3:04 PM
> To: Ian Cooper
> Cc: Christian Kuhtz; cdn@ops.ietf.org
> Subject: Re: new to cdn internetworking
>
>
> On Tue, Nov 13, 2001 at 12:00:29PM -0800, Ian Cooper wrote:
> [..]
> > Is the decommissioning of proxies something that's limited to North
> > America, or is this something you're seeing happening on an
> international
> > basis?
>
> The variables that drive the decommissioning of proxies apply
> globally from everything I've seen.  The trend is steady as well.
>
> The thresholds in North America tend to be a little
> different.  However,
> the problems, issues, constraints etc are very similiar in NA vs ROW.
>
> One could argue that it is only a matter of time that this will happen
> overseas as well.  Given the rate at which this is observed,
> the changes
> have been so rapid that overseas should feels this in the fairly
> immediate future.
>
>

Your argument has major holes in it. The cost of bandwidth is going
down, yes. How do you know the cost of local caching won't go down
next year?

Basing long term standards and protocols on short-term relative
pricing shifts is not good planning.

Inherent laws of physics say that caching material closer to the
end consumer is cost effective when there are multiple hits.
Factors that could make local caching more viable include:

a)      Higher broadband penetration. Long-haul lines feeding dial-up
     customers are one thing, customers with fat pipes that want
     music and/or video streamed is another.
b)   Potential for "mass appeal" content -- such as music and video
     would increase the "hit" rate.
c)   Municipal networks could expand the realm of what is "local"
     allowing caches to operate over a wider audience and still
     have cheap local delivery.
d)   A content networking standard would reduce the overhead of
     managing edge caches -- thereby making them more cost effective.

Standards should be technology neutral, not trying to declare certain
options to be dead.