As promised, and before more time goes by, the wrap-up on the highlights
from the O'Reilly p2p conference.
Without further, here are some of my notes on some of the sessions,
thoughts, summaries, quotes, etc. etc. as I could best make them from my
own notes! ;-)
Many thanks to David Wiley (DW), Ph.D, from Utah State University, and
Steve Pugh (SP), Sr. Network Engineer at UT, who also contributed with
insights, memory refreshes, and their own comments and impressions on the
conference.
This is the FIRST of a series of messages that contain the
hightlights. The very last one has our very own impressions. Please feel
free to comment as well. Enjoy! --ana
The O'Reilly P2P Conference -- San Francisco, CA. Feb. 14 - 16, 2001
On this email:
- Words from Tim O'Reilly about p2p
- Lessons from Napster
- P2P in the enterprise
2/14/01:
Welcome: Tim O'Reilly
- "What Napster showed us is the shape of the next-generation
Internet." "Napster has crystallized the notion that we are entering the
network era."
- With the Napster buzz, along came the erroneous message that p2p is
about copyright issues and piracy, when "it is really about sharing,
communication and collaboration."
DW: Napster is helping more and more people see that our current economic
model just doesn't fit the information age. The current scarcity model,
where supply and demand are forces that shape the economy, has been
obviated. Digital resources are not scarce, there is an infinite supply
(assuming that bandwidth is infinite over time), and the notion of
supply and demand (when there's a large supply, prices are low; when
there's a limited supply, prices are high) disappears. What we see with
the RIAA and MPAA lawsuits are two scarcity-based business models
refusing to see reality. Until they do, they'll continue trying to
create "artificial scarcity" through technological or legal means in
order to protect their 20th century business model. Too bad they won't
spend 1/2 their legal budget addressing the 21st century.
********************************************************************
Lessons from Napster: Clay Shirky, author and partner with The Accelerator
Group.
- The P2P concept: "aggregating resources at the edges of the
network." The possibilities for machines at the edge of the network are
huge: there are 10 quadrillion computing cycles and 10,000 terabytes of
storage space available on PCs connected to the Internet."
- 4 BIG lessons from Napster:
1. Napster is decentralized enough to qualify as a p2p application. While
the "centralized" aspect of it handles the "master" databases, it is
decentralized in the sense that each user can have lists of people and
music they are willing to use their resources on.
2. Napster's poor interface did not interfere with enabling a system that
allowed users to do something they want to do. Mentioned how Napster made
it easy to hide the complexities of the Internet (e.g., setting up a
domain name server and the tedious steps one has to go through). Napster
offers something that anybody can use and that works.
3. Napster enabled people to "sneak p2p applications in under their
network administrators." The joys of subversion! (as when in the 80s
people were sneaking their own PCs into their work under the noses of
mainframe administrators).
4. Beware of the "ominous centralization represented by the databases at
AOL Instant Messenger and Napster." P2P applications as these set up
their own name service ý beware of those in control of those
databases. "...P2P has brought back communication and power to the people
but it has done this by creating private address spaces. The risk cannot
be overstated. Resist this. This is the biggest opportunity and the
biggest challenge for this group."
(On a related note, check this Salon's article:
http://www.salon.com/tech/feature/2001/02/09/napster_parasites/index.html)
DW: While private namespacing certainly *is* an issue, search services in
any p2p app must remain centralized for the near future, as demonstrated
by Jordan Ritter's recent analysis of Gnutella, which employs a truly
distributed search. Speaking of the amount of bandwidth necessary to
support search requests and responses (not downloading any files, mind
you!) he concludes: "From the charts above, it becomes mind-numbingly
clear that the Gnutella distributed architecture is fundamentally flawed
and can have a horrific impact on any network. On a slow day, a
GnutellaNet would have to move 2.4 gigabytes per second in order to
support numbers of users comparable to Napster. On a heavy day, 8
gigabytes per second." Remember, this is only query traffic - not what
happens when people actually start up/downloading files.
(http://www.monkey.org/~dugsong/mirror/gnutella.html)
- Closing remarks: The Network is the computer
"50 years ago, Thomas Watson estimated there was a worldwide need for
maybe five computers. We now know that that number was wrong. He
overestimated by 4"
********************************************************************
P2P in the enterprise: panel discussion
Ray Ozzie (Groove Networks), Eric Schmidt (Novell), Jonathan Hare
(Consilient), Andrew Chen (Entropia), Erick Von Schweber
(Cacheon); moderator: Bill Bunham (Softbank Ventures)
- The big question: how will p2p models make money?
- Is P2P this year's "push"-like or hand-based computing-hype? Consensus
that those technologies sold themselves too soon
- P2P deterrents:
- p2p's questionable reputation: no respect for my IP, like Napster;
- a bandwidth hog like Gnutella and Napster, chaotic and out of
control relative to servers
- anxiety about: new models, unknown infrastructure impact; viruses
sneaking in, etc
- PC lock-down and mobile code policies
- Central control expectations:
- Network: bandwidth use, firewall and proxy use
- security: identity, key management, access control, mobile code
- behavior of the applications
***************
If you care for more, go to
http://www.openp2p.com/pub/a/p2p/conference/index.html
for more in-depth and specifics about the conference.
|