Hari
Balakrishnan Assistant Professor of Electrical Engineering
and Computer Science (EECS)
"You're sitting at work and you want to know if you've closed the
garage door. How do you find the two or three cameras monitoring
your garage, among the many devices in your home?" asks Hari
Balakrishnan.
"This is the kind of question we're likely to
pose in the next five to ten years, he says, as we start linking
everything from cameras to coffee machines to the Internet. Networks
of the near future will enable us to run home appliances, monitor
office equipment, gather information from remote sensors, and
operate many other devices.
All of this will require careful
planning. Balakrishnan and his research group of ten students are
now developing ways to effectively support networks of devices (in
addition to computers) on the Internet, such as those envisioned by
the Oxygen
project. The team aims to keep the Internet functioning smoothly
as devices generating video, audio, sensor readings, and other data
formats are added on. "Our problem is to ensure that we can still
have a good network architecture, get good performance, and have a
robust system," says Balakrishnan. Towards that end, his research
group is tackling two major projects -- Wireless Networks of Devices
(WIND) and the Congestion
Manager (CM).
WIND is
a system of middleware and protocols that will support new
applications in dynamic, mobile networks of devices, sensors,
actuators, and computers. Its primary design goal is to make
networks self-configuring and applications self-organizing, so they
can discover services and function without manual intervention. WIND
project research areas include automatic network configuration in ad
hoc environments, packet routing, adaptation to changing network
conditions, security/privacy problems, and resource discovery. To
enable automatic discovery and use of computers, devices, and other
resources in dynamic, mobile networks, the group has designed an
innovative "Intentional Naming System."
As a host of new
devices and applications get networked, the Congestion Manager will
be used to monitor and optimize Internet performance. Unlike the
standard protocol TCP, the CM can efficiently handle increasing
Internet congestion due to audio, video, and other real-time
streaming applications, as well as from multiple short Web
connections frequently launched by Web browsers. Preliminary results
from simulations of the CM architecture show its effectiveness in
enabling networks to expeditiously manage Internet congestion from
these sources, and applications to easily adapt to changing network
conditions. A prototype is now in development.
"I think life
will become easier if a lot of things in our daily physical world
are networked," says Balakrishnan, who joined LCS in September 1998.
He became an EECS Assistant Professor after completing a PhD thesis
on reliable data transport over wireless networks at the University
of California, Berkeley's EECS department. The thesis garnered the
Association for Computing Machinery's doctoral dissertation award.
Balakrishnan currently teaches Computer Networks (6.892) and
Computer Systems Engineering (6.033).
|