Every once in a while I get to see a quote from Einstein, in someone's
e-mail signature block, on some mailing list, that goes something like this:
'Problems cannot be solved at the same level of thinking that created them.'
To me this quote perfectly describes our predicament with P2P. The "How
much bandwidth is reasonable?" thread started out as a simple question from
Joshua Wright who has a job to do. He has to figure out how much Bw is
reasonable per user in a residence facility (for design purposes I assume).
Einstein tells me that whatever Joshua comes up with, he will be way off !!
Sorry Joshua :-( He, like others, will come back after Easter or after
summer holidays or after Christmas and have major out-of-control fires on
his hands. One *cannot* know ahead of time that usage will e.g. double over
a certain, however relatively small, delta t. But you should bet on this -
it has hppened and it will happen again.
In the last couple of days, in addition to Napster, we have heard of the
advent of "legitimate" P2P uses of the Internet. (Jerry let's hold on to
the gasoline for now ;-) ).We've heard about learnster, the educommons
project, and the well known [log in to unmask] In this last month I've heard of
many new P2P projects, more than I can clearly recall. As examples, Wall
Street (http://www.canarie.ca/MLISTS/testnet2001/0039.html) wants to do P2P,
as do a growing group of North American high schools participating in a
project to do distributed cosmic radiation detection and analysis, etc. Has
P2P really arrived yet?
Historically Universities have been the birthplaces of new movements, new
rising sentiment, new trends (social, technological, or other), new
paradigms, etc. which eventually spread out into the layers of society. I
think it is no different now. he phenomenon we have been seeing in campus
networks may just be an early sign of what is about to happen in the
commodity Internet. What happens when P2P breaks loose in all spheres of
activity (legit or not)? ADSL and shared cable-modem lines are still being
rolled out. How quickly might they become obsolete?
A good article (http://www.canarie.ca/MLISTS/testnet2001/0056.html and
forwarded on the P2P list as well) says that "if p2p apps become ubiquitous,
they could break the existing business models of many Internet Service
Providers and force them to raise their prices " Yes - just what we all
wanted to hear!
I think Joe and Jerry and others have touched upon what might be done to
solve the problem (from a higher level than created it). Settlement free
exchange-point based peering, massive peering, GigaPOPs, dark fiber nets,
neutral colo facilities, self organizing nets, etc.
Anyway I have thoroughly enjoyed this thread. Thanks
:-)
|