 |
|
DAILY NEWS AND INFORMATION
FOR THE GLOBAL GRID COMMUNITY / JULY 28, 2003; VOL. 2 NO. 30
|
Special Features:
PROFESSOR DEVELOPS GRID USING PEER TO PEER SYSTEMS
With music file sharing receiving loads of attention in the media because
of
proposed lawsuits by music companies, the fact that the peer to peer networks
that make the sharing possible can be used for other purposes is being
lost.
Y. Charlie Hu, associate professor of electrical and computer engineering,
hopes to change that.
Hu recently received a multi-hundred thousand dollar grant from the
National
Science Foundation, which he is putting toward research on improving computer
networking.
As it stands today, computer networks are restricted in reliability and
stability by the fact that a network must have a single manager computer, or
node, to oversee the network, however, if that manager fails the whole system
goes down.
Hu hopes to use the technology that is used in peer to peer networking to
create a more stable server where all computers can share with each other
freely as opposed to many systems client-server models. Hu said that the
technology that is used to download music is much more powerful than what it
is being utilized for because until recently it wasn't believed it could be
applied to the same functions that the single node network is utilized
for.
"This was the first proposal to applying peer to peer to grid computing,"
Hu
said. "Up until this the two have been totally independent."
While peer to peer was used for file sharing the grid computing was used to
link supercomputers for scientific purposes. The problem with the grid system
is that it is too centralized in case of problems or heavy traffic, Hu
believes that using peer to peer technology will allow computers to
decentralize.
Basically, Hu's proposed system would allow networks to have a group of
nodes
working as one to solve problems and run the networks.
Hu said the future of networking lies in the technology he is researching,
which is why he believes the National Science Foundation gave him the research
grant.
The grant was part of the Faculty Early Career Development program that the
foundation has set up in order to get young scientists research experience.
Hu, who has received a doctorate degree and has been at Purdue for about a
year and a half, said that most grants go to the person with the most
experience, which can make it difficult for recent graduates to get grants
over researchers with 20-plus years of experience, but this one is
different.
This grant can only be won by researchers just out of school and can only
be
won for one time. Hu was one of three Purdue researchers to receive the grant
and one of more than 2,000 applicants for about 350 grants. The other Purdue
faculty members to receive the four or five year grant were Phillip Dunston,
assistant professor of civil engineering, and Kendall Thomson, assistant
professor of chemical engineering.
The National Science foundation is a federal program funded by tax money
that
is responsible for managing research grants for both academic and business
fields. The foundation, comprised of educators from around the nation, meets
to decide on a research focus, review applications from researchers and assign
grants, Hu said.
"They draft a proposal for what they are looking for then they announce
that
proposal and people apply for grants in that area of specialization," Hu
said.
|