[e2e] Re: Network expansion based on customer demand (was: Re: [e2e] Fundamental Questions about Router Queue in High Speed IP Networks)

[David P. Reed]

At 01:04 PM 8/24/2001 +0200, Anders Bergsten wrote:
>But, would you say that we (as an ISP for example) have enough
>information to judge where to expand the network to meet the demand that
>the customers are willing to pay for?
>
>As I see it, we only have information about the resource utilisation on
>single links to judge the demand.

I think you nailed the key question here.

Of course the links are the wrong place to measure demand, since the 
"end-to-end" demand is for short-latency capacity from endpoint to 
endpoint.  The fact that it gets aggregated on a particular link is an 
artifact mostly of routing choice...  and routing choice is only loosely 
connected to demand (especially unsatisfied demand that TCP is holding back).

There are very useful signals outside the TCP stacks (perhaps in the user's 
mind only) about demand.

One way to think about this is to find a way to communicate a user's 
willingness to pay, into a "market" for future capacity.  I've thought 
about this a fair amount, and I think there's a simple scheme that could be 
made to work.  But as a self-employed person - well, there's not a lot of 
funding for system-level paradigm shifts out there (maybe there will be now 
that the router companies can't get customers by loaning them the money to 
buy their products).

--------------------------------------------
WWW Page: http://www.reed.com/dpr.html