It all began last year as a kid's simple effort to get new
music. After Shawn Fanning, then 18, wrote a program enabling him and his
friends to share tunes, Napster was born and Fanning became a hero. Thanks to a
multimillion-dollar lawsuit, the young music lover's pet project has changed,
but his early experiment with peer-to-peer computing has also altered the Internet.
Just as B2C and B2B became buzzwords, everyone from
entrepreneurs to venture capitalists is touting P2P as the next big thing.
Statistics from PricewaterhouseCoopers and Forrester Research show that P2P
technologies have transformed the recording industry, but little has been said
about how distributed computing might affect companies in other industries, and
how new applications of distributed computing could affect the Internet economy
as a whole.
How can companies use the technology to improve
communication with customers and ameliorate customer service? How can
e-business companies incorporate the technology into their business models?
These may be the questions of the future.
"Peer-to-peer technology will inevitably spawn many
companies that use Napster-like services to distribute product," says Clay
Shirky, a partner at the venture-capital firm The Accelerator Group. "What
distinguishes these companies will be the way they understand that simply using
this technology isn't going to change anything unless there's a compelling
reason to do so."
Perhaps the most talked-about firm employing
technology-distributed computing in its infrastructure is Groove Networks Inc.,
based in Beverly, Massachusetts, and run by Ray Ozzie, the inventor of Lotus
Notes. The company's model is simple: Develop interactive spaces, facilitate
collaboration within those spaces, and run it all off of wasted computing power
on the participating machines. At the center of this approach is a proprietary
service that companies employ to add particular functionalities to their
pre-existing management systems including chat, instant messaging, video or
audio.
Raghu Ramakrishnan, founder and CTO of Quiq Inc., based in
San Mateo, California, was so impressed with the model that he used it as a
guide to design his business.
At its core, Quiq and Groove are very similar. Through
Quiq's flagship service, companies will be able to connect with their partners,
users and vendors, and form communities that allow each to interact and share
data accordingly. While Groove applies to customers of every ilk, Quiq's
service will be only for those who meet specific qualifications. Still,
according to Kevin Lenzo, assistant scientist at Carnegie Mellon University's
Institute for Software Research, International, these technologies are only the
beginning.
"In terms of communication, I see distributed computing
as the spontaneous formation of communities, irrespective of geography,"
he says. "The technology allows for the kind of communitarian behavior
that wouldn't necessarily happen otherwise on the Internet if you just had
random people logging on to a central place."
While Groove seeks to aid employee communication and
customer relationships, other companies are looking to incorporate P2P
technology into the very infrastructure that drives their e-businesses. Some are
building networks that put e-commerce around the people who use them most:
customers.
Startup companies trying to utilize this tactic include
MyCIO.com, which recently rolled out a Napster-like P2P file sharing technology
and service for managing anti-virus updates, and InfraSearch, which uses P2P to
develop searches that connect buyers directly with sellers and vice versa. Even
Internet powerhouse eBay admits that it's exploring the feasibility of
incorporating some sort of a distributed model into the company's now-famous
marketplace.
Although these firms differ by industry, the idea behind
deploying P2P is the same: Decentralize the infrastructure in an attempt to
lower costs and improve efficiency. At MyCIO, for example, the first five users
in a network automatically retrieve anti-virus updates, then act as servers for
everyone else. Every time one of these users gets a new update, the machine
sends out queries to see which other users need it, then supplies those who do
with a copy of the file. Zach Nelson, the company's president and CEO, says
that because the computers communicate with each other, there is no need for a
client-side server to coordinate the updates.
"We saw that the whole client/server model was too
slow," he explains. "If you're shipping bits and bytes of anything,
when you're talking about large package distribution to a number of customers,
peer-to-peer is the only way to go."
But the business-to-consumer venue isn't the only one for
distributed computing. In an essay published recently by the Harvard Business
Review, Andrew McAfee writes that those in the business-to-business market can
use P2P to avoid membership fees, reduce networking complexity on the back-end,
and coordinate with companies in a variety of industries, rather than sticking
to affiliations with companies just like theirs.
"There are many uses we can make of intelligent
machines, but the key to figuring out what they are is identifying that all
interfaces are different," says Dan Bricklin, founder and CTO of software
company Trellix Corp. and creator of VisiCalc, the first electronic
spreadsheet. "If you use the model of 'choose the right tools,' you'll
hopefully be successful. If you copy someone else without thinking, you will
fail."
According to venture capitalist Shirky, the real answer may
be a melange of decentralized and centralized computing that incorporates P2P
and client/server at the same time. He insists that this approach would
incorporate the best of both worlds: the functionality and scalability of a
distributed infrastructure with the security and reliability of a centralized
one.
"The magic P2P pixie powder won't last forever,"
Shirky opines. "For this model to take hold in the e-commerce space, we've
got to see some diversity."
Matt Villano is a contributor to The Next Big Thing. Send email.
Next: A P2P primer