Search
Sites and Affiliations
Leadership
Research and Applications
Major Accomplishments
Frequently Asked Questions
Knowledge and Technology Transfer
Calendar
Education and Outreach
Media Resources
Technical Reports and Publications
Parallel Computing Research - Our Quarterly Newsletter
Contact Information
CRPC Home Page

UniGuide Featured Site
BANDWIDTH BUSTER

Rice professor leads triumvirate of Houston universities into uncharted cyberspace

Source: Houston Citynews, Wednesday, July 30, 1997
By Dave Sarles
Associate Editor

Forget rivalries, Houston universities have joined forces to design the next Internet.

The Houston Area Computational Science Consortium (HACSC) - UH, Rice, Baylor College of Medicine - plans to incorporate high performance computing and virtual reality into the next Internet. The research that the universities are focusing on is increasing the bandwidth capabilities so that full-motion video and new applications can be transferred in real-time.

And at the helm of this group is Rice computer science Professor Ken Kennedy.

These changes will metamorphosize today's Internet. So much that you might not recognize it in the next decade or so. It may look like truly interactive TV, where everyone has the opportunity to have their own channel.

"In many ways it will be the same, but it will also be a very different experience for users," Kennedy told Citynews in a recent interview on the groundbreaking project. "It will probably be more like what is on a CD ROM today. But downloaded over the Internet."

The three Houston schools are integrally linked to this sequel. A sequel that's got a higher budget than any Hollywood movie - some $500 million.

The government is investing that money into research and development of the Next Generation Internet. President Clinton handpicked Ken Kennedy to co-chair the Advisory Committee on High Performance Computing and Communications, Information Technology and the Next Generation Internet.

"It is a great honor and an opportunity," Kennedy says. "The most exciting part is being in on the formation of new initiatives like NGI that have the potential to change everything."

NGI and the Internet II are two different entities that are often confused. Kennedy says it's important to understand the differences and also how they relate to each other.

"Internet II is a consortium of US universities that have joined together to establish high-speed connections to one another," he says. "It has no formal relationship to the Next Generation Internet, which is a federal research initiative to establish the technology infrastructure that will be needed to bring the current Internet to bandwidths (amount of information that can be transferred in a second) to 100 or even 1000 times what it is today.

"The Internet II consortium may benefit from the NGI's proposed support for a testbed that would interconnect 100 or more research labs, especially those at universities. However, the NGI support would pay only part of the cost of establishing connections that are needed for the testbed, significant costs would fall on the universities themselves."

Kennedy and other academics from across the country are doing research that will explode the current bandwidth to unheard of heights. This can be done thanks to a predecessor program started by the National Science Foundation.

The vBNS program (which stands for very-high-speed Backbone Network Service) has provided partial support for connecting universities to an experimental network that achieved bandwidths that are initially three times faster than typically available to universities, but later will rise to 12 to 24 times faster. Since the vBNS technology is being used for Internet II, the three universities, which have also joined I2, have an early connection to the network being developed, according to Kennedy.

"Currently we are still connected at our previous speeds of 45 megabits per second while we work out the details of getting the higher bandwidth connections," he says. "In the meantime we have established partnerships with major research computing sites, especially the newly-funded Partnerships for Advanced Computational Infrastructure that have been established at UC San Diego and University of Illinois at Urbana-Champaign.

"A major goal of our collaborations with these partnerships is to develop software that will support truly distributed applications - those that run on a collection of high-speed computers located around the nation. Rice, UH and BMC will be a part of national testbeds for experimenting with such distributed applications, which I believe are the wave of the future."

Estimates say that in a few years we may be able to transfer 2.2 trillion bits over the Internet, that's 1000 times what it is today. Today's average home transfer rate of 28 thousand bits will seem archaic in comparison.

"The long-term idea is to bring the current Internet up to support higher bandwidths at every link, including the links to the home, which will soon be over a megabit per second - nearly 50 times the typical access speeds available to the typical home today."

The Internet II is not unlike what the original Internet was before commercialization engulfed it. It was a group of universities linked together over a huge computer network.

"The most important thing about the Next Generation Internet is that it will enable the development of new kinds of applications, some of which we cannot envision today. It is clear that photo-realistic images and full-motion video will be a part of such applications," Kennedy says.

"Applications close to home involve things like home shopping, where buyers can use virtual environments on their home PCs to explore the products they are thinking about buying. This could even involve performing simple tests on a model."

While over at the University of Houston, computer science professor Lenart Johnsson is working on how the Houston universities will connect to this new Internet.

"We are working on creating a gigapop here in Houston. That's where the messages come in and out over the computer network," Johnsson says. "It's not unlike phone switches that the telephone companies have. Gigapops are doing the same thing that those do, but on a much grander scale.

"We are all working together to come up with the plan to establish one of these gigapops that will be the main connection for the Houston schools to the Internet II. That has positive consequences for all the educational institutes in this state."

UH is also home to The Cave - a virtual reality environment that is unique. It will also be a place where research is done to incorporate VR into the NGI.

"For very complex data sets that aren't easy to understand in a 2-D environment, virtual reality creates an immersive environment. When dealing with a flat screen you can't really get a feel for active data that needs to be visualized

"UH has one of the most powerful computers in the state, so in terms of the Internet II, we're working on ways to link virtual environments in more than one place by using a high-speed, high-bandwidth connection. This network is the key component in making this form of collaborative research happen across the country and across the world."

An intriguing concept about the future capacity of the Internet was brought up by Raj Reddy, Carnegie Mellon University Dean of computer science and member of Dr. Kennedy's panel. He says that "a perabyte (1015 bytes) of information for $100 - that's equivalent to all the printed material that's ever been created."

"Raj is just saying that we will be able to afford large amounts of storage on our own PCs in the not-too-distant future," Kennedy says. "But what will we put there? And how will we decide what information from the vast NGI resource is useful and what is unreliable - that is a major challenge for future research."

The future is exactly what Professor Kennedy is ironing out in his committee. The who, what, where and when of the NGI is important and a central part of this country's future.

"There's a problem with networking - we must be careful about placement of the gigapops. We don't want to create technology-haves and technology-have-nots," Kennedy says.

"The Federal program has to think of ways to work this out. They don't want to make it too expensive for the schools to network. Right now, Houston has an edge because we're going to get a gigapop.

"But the whole infrastructure issue is complex. It might be cheaper to lay two lines, one for data and one for phone service. Given the demand for networking in this country, we are going to have to be ready to upgrade the infrastructure constantly and we will always be in the process of improving it.

"There's going to have to be a lot of research done to make this all work. There are several problems that haven't been solved but we have partial solutions and the initiative to work them out."


Sites & Affiliations | Leadership | Research & Applications | Major Accomplishments | FAQ | Search | Knowledge & Technology Transfer | Calendar of Events | Education & Outreach | Media Resources | Technical Reports & Publications | Parallel Computing Research Quarterly Newsletter | News Archives | Contact Information


Hipersoft | CRPC