In the mid-1800s, Abilene, KS was the official end of the Chisholm Trail. Today, Abilene is the name of the network that may become our next Internet. Abilene may prove to be to network communication what the steam locomotive was to stagecoach travel. The new line travels faster than the "old" one while visiting some new and exciting territories. Although Abilene isn't brand new (planning for it began in 1995), so far, its only scheduled stops are on university campuses and other research hotbeds. In a few years, the public may be invited to hop onto Abilene. For now, we can marvel from afar at Abilene's speed and at the robust applications that run along the fast new line.xml:namespace prefix = o ns = "urn:schemas-microsoft-com:office:office" />

Abilene Isn't Internet2

Abilene is the name given to the advanced backbone network that connects regional networks and network aggregation points called gigaPoPs. Abilene was developed by the University Corp. for Advanced Internet Development (UCAID; www.abilene.org) to connect the people and places of Internet2 with each other as well as with peer networks worldwide.

            Internet2 (www.internet2.edu) is a nonprofit consortium of about 185 universities and research centers working in conjunction with government researchers from the National Science Foundation (NSF; www.nsf.gov), Next Generation Internet (NGI; www.ngi.gov), and other advanced networking initiatives worldwide. While NGI and Internet2 work on similar projects and sometimes share representation in technical organizations, they maintain separate roles and goals. NGI is a federal initiative with goals of developing and delivering advanced networking technologies to the commercial Internet, while Internet2 is dedicated to developing and providing services for academia first—universities and high schools today, museums and libraries to be included soon.

            If Internet2's relatively short history sounds familiar, it's because it has followed a development path similar to that of today's commercial Internet. In the 1980s, the NSF created NSFNET, a high-speed (for the time) network backbone linking a group of regional networks. The commercial Internet grew out of these efforts, and NSFNET was decommissioned in 1995. Since then, many Internet2 institutions have received funding from NSF to, once again, create a high-speed network to rapidly transfer new network services and applications to educational facilities and, ultimately, to the public.

            Meanwhile, NSF participates in several separate but coordinated programs designed to make future Internet communications faster and better. Last year, about 150 Internet2 member universities received grants to support connections not just to Abilene, but to the very high-performance Backbone Network Service (vBNS), developed by NSF and MCI/WorldCom to support research and education. The Internet2 group is working on what very likely will be the next incarnation of our commercial Internet.

            But, Internet2 is not designed for commercial applications at this time. It's dedicated to research and educational endeavors, with the understanding that what works on Abilene today will likely find commercial acceptance sooner rather than later. While Internet2's high-performance networks are separate from the commercial Web, there are some points at local and regional levels where universities use network links from commercial Internet service providers to connect to the Internet2 backbone, says Greg Wood, director of communications for Internet2. The two networks are largely compatible, he says, but there are significant differences between the new technology and that of the current commercial Internet.

            Abilene's most obvious advantage is its speed. Like the vBNS, Abilene was built to move at 2.4 Gbps, which is about 45,000 times faster than a standard modem. Now, Abilene is being upgraded to 9.6 Gbps, which is 200,000 times faster than a standard modem. The upgrade is expected to be completed in 2003.

            In addition to increased speed, other intriguing Abilene attributes include being able to prevent data loss and minimize transmission delays by prioritizing data so it travels without interruptions, a function referred to as Quality of Service (QoS). Initial QoS testing has been positive, and one day soon it may prove good enough for surgeons to use to operate on patients via remote access and remote instrumentation.

            Another tantalizing feature of Abilene's technology is multicasting, which allows users to send a single stream from one computer to hundreds or even thousands of others; rather than sending hundreds or thousands of copies, only one is sent. The single stream essentially duplicates itself, splitting off copies to send out. Multicasting is attractive because it avoids bottlenecks and allows for faster transmission over average bandwidth.

            Internet2's real-time collaboration environments will allow for long-distance, high-quality, multisite videoconferencing for academic researchers and R&D departments. The same high-quality video (think live, interactive television) plus remote access to resources such as telescopes and microscopes located all over the world will provide great benefits for scientists. It shouldn't come as a surprise, then, that the Internet2 consortium has major university members in all 50 states, and high schools are now signing on to participate.

            A few businesses have signed on as Internet2 sponsors, ranging from Cisco, Nortel, and Qwest to Ford and Johnson & Johnson. Cisco was the first corporate sponsor, working with Internet2 participants on network design and donating millions of dollars worth of equipment to member universities. Nortel has been a major participant too, working on QoS and other network issues seen as key to the success of future Internet traffic. The equipment donated by Cisco, Nortel, and Qwest has been estimated to be worth hundreds of millions of dollars.

            Denver-based Qwest struck what can only be called a win-win agreement with Internet2 back in 1998. The then-relatively unknown company invested about $500 million to get involved with a high-speed network and the research that Abilene would carry. "It gave us credibility," says John Walker, Qwest's vice president of government and education solutions. When Internet2 approached Qwest last summer, the company was still interested, Walker explains, because, "By that time, most of the commercial ISPs had caught up to the technology that Abilene used." A new investment (about $300 million this time) gave Qwest the chance to get ahead of the commercial curve by upgrading Abilene to run at about 10 Gbps. "Essentially, Abilene is a subportion of our network," Walker explains. "We own the fiber that it runs on and Internet2 owns the equipment and routers."

            If Internet2 is the future of the Web as we know it—and it probably is—the obvious question to ask is, what's happening on Abilene now?

Some Applications and Research Projects

Thanks to Abilene's high-speed transmission, videoconferencing from the desktop is finally a worthwhile endeavor. What was cost-prohibitive to do in person is now feasible with Internet2, and examples abound. Julliard master musicians can visit schools in Idaho without spending time or money on travel; deaf students can hold live, face-to-face debates with students at schools across the country.

            Last fall, Internet2 announced a plan called the K20 Initiative, which will work to bring the high-performance research and education network previously reserved for universities to high schools and community colleges. Imagine your son's science class gaining remote use of a scanning electron microscope to analyze samples collected (remotely and in real time) from the ocean floor. Your daughter's Spanish class in Topeka can have a joint class with students in Madrid, and both classes can tune in for virtual music lessons with the first horn from the Cleveland Symphony Orchestra. Abilene technology and Internet2 partnerships can make such things happen today.

            Ford signed up as a corporate sponsor of Internet2 at the end of 2001, probably lured by the prospect of enhancing collaboration and development efforts among geographically scattered work teams. Ford's first project will focus on multisite, television-quality videoconferencing. Undoubtedly, the prospect of collaboration among designers and engineers physically separated by many miles will lure other companies to sign on in the near future.

            The three common types of videoconferencing, H.323, MPEG-2, and Access Grid, each have their own proponents and strengths, and the various technologies can be combined to some extent for greater effect. In general, MPEG-2 offers higher bandwidth at a higher cost than the other technologies; Access Grid technology is especially well suited for multipoint conferencing. (Access Grid was used at last year's supercomputing conference in Denver to host a performance of dancers and musicians in four different locations.) H.323 (Internet) videoconferencing is rapidly replacing H.320 (ISDN telephone) videoconferencing. "It's the tool that everyone is going to have when we replace our telephones with videophones," says Bob Dixon.

            Dixon holds the same title, chief research engineer, at two places: at OARnet, an academic Internet service provider in Ohio that is also an Internet2 gigaPoP location, and at the office of the CIO at Ohio State University. He also holds the unofficial title of Mr. H.323. Last summer, Dixon and other H.323 researchers engineered (and participated in) a barbershop quartet in which the four singers performed in four different time zones. The conductor (Dixon) and audience were in a fifth location. While the "maestro" says the quartet could have benefited from a bit more practice, the demonstration of H.323 technology got rave reviews.

            Videophones that use H.323 technology are commercially available in the $500 price range and can deliver good results over T1 lines, says Dixon. For now, the masses aren't rushing to replace their standard household telephones, but a relative of the videophone is increasingly appearing in hospital operating rooms.

Telemedicine

Videoconferencing in small or rural hospitals may soon offer tremendous benefits to critically ill or injured patients. Today, Dixon explains, if a patient is opened up and the operating surgeon realizes that procedures are required that are beyond his or her expertise, the patient is quickly sewn back up and transported to a larger hospital where a specialist resumes the operation. With a video link in place, the attending surgeon could consult with the specialist, who would essentially provide over-the-shoulder consultation; no physical transportation would be required.

            Dixon says this is proven technology, not conjecture. "At the moment, the problem is not a technological one," he says, explaining that the roadblocks remaining to telemedicine are primarily legal. Questions regarding who pays for teleconsulting services and how insurance and malpractice entities will consider such procedures are broad in scope and difficult to manage. "The long-term hope is that it could be used for remote consultations," says Dixon, who has seen the possibilities up close and remotely, watching laparoscopic surgeries.

            The minimally invasive techniques of laparoscopic surgery make for great video. Before the surgery, several very small incisions are made in the patient's abdomen. Tiny cameras and lights are inserted into the openings. Instead of looking at the patient to perform the surgery, the surgeon guides his hands while watching onscreen images from the inserted cameras. Students in medical schools can benefit from watching the procedure in real time and surgeries can be archived for future viewing.

            Laparoscopic surgeries aren't the only situations in which telemedicine can have tremendous value in teaching situations. The Visible Human Project (www.nlm.nih.gov/research/visible/visible_human.html), begun in the early 1990s by a team of Michigan researchers, has created two digital cadavers (one male, one female) on which medical students can operate again and again. Using modeling software, the Visible Human Project allows medical researchers to create virtual injuries on the body, so as to study prevention and cure.

Virtual Collaboration, Tours, and Concerts

When technology such as the ImmersaDesk is used with videoconferencing technology, the result is almost as good as being there. The ImmersaDesk looks like a drafting table. Users wear liquid-crystal shutter glasses and interact with images on a rear-projected screen using a wand, voice commands, or both. The ImmersaDesk isn't new (it was developed by the Electronic Visualization Lab at University of Illinois at Chicago and has been available since 1995), but when paired with the fastest network on earth, its potential expands enormously. While virtual-reality technology has always appealed to game developers, it has many fans in the less glamorous but equally creative world of mechanical research and development. Rapid prototyping in a 3D space holds obvious appeal for engineers and designers.

            The union of equipment such as the ImmersaDesk, virtual-reality applications, and Internet2 bandwidth produces a number of new options for other users, such as museum tour guides. Anne Doyle, Internet2 manager for arts and humanities, toured the Spirited Ruins created by Boston University as a demonstration of Internet2's possibilities in last summer's Virtual Arts Festival. In the tour, visitors in Ann Arbor were led through an artistic interpretation of an ancient palace by a tour guide in Boston.

            In addition to experiencing art in a lively and interactive way, there are other disciplines making use of this technology, Doyle says. Using the same concept, for example, an architectural specialist could virtually tour and make structural suggestions for repairing a building damaged by an earthquake. Architectural consultations with experts worldwide might become more feasible when they're traveling via fiber optic cables.

Peer-to-Peer Applications

New Protocol, Infinite Internet Addresses?

The current protocol used to assign addresses to devices on the commercial Internet is the 32-bit IPv4. IPv4 has a finite number of addresses, and our love of phones, Palm devices, and all things digitally connected has begun to push the limit. The 128-bit IPv6 offers essentially unlimited IP addresses, enough to quench our thirst for current portable technology and probably for the remote-control computers that will govern our household appliances one day.

            Walker says that while Qwest has not developed any commercial products as a direct result of formal ties with Internet2, some technologies are developing on parallel tracks—and IPv6 is one example of where this might occur. Qwest IP engineers who are working on IPv6 meet occasionally with Abilene engineers, Walker says, and did so before the official Internet2 IPv6 research lab was operational. The hope is that IPv6 will be ready for all Internet users before we reach the limits of IPv4. (For more on IPv6, see The Internet on page 68.)

When There's a Whole Lot of Data Going On

The National Digital Mammography Archive is a collaborative effort of the National Library of Medicine and at least five universities. Its aim is to support traditional breast-cancer screening through maintaining and distributing prior breast-examination data and family medical history to provide for interpretation and expert consultations and to create educational and training tools for diagnosis and treatment. To put it mildly, that's a lot of information. Abilene makes short work of mining, managing, and analyzing such vast amounts of data.

            Another place where this has proven true is at the Center for Analysis and Prediction of Storms (CAPS). CAPS was established at the University of Oklahoma in 1989 as one of the first 11 National Science Foundation Science and Technology Centers, and it has recently licensed its weather prediction technology to a local private company, Weather Decisions Technologies, which does weather-impact studies for some large customers, including major commercial airlines.

            CAPS' continuing research was a natural fit for an Internet2 initiative, allowing scientists to use ever-expanding collections of data in their research, and allowing the public to benefit from better predictions, thanks to a new system for the real-time compression and Internet-based transmission of high-volume data from the nationwide network of NEXRAD Doppler radars. The combination of research and speedy transmission may benefit us all through the accurate prediction of short-term hazardous weather, giving those in the way time to prepare for it.

            The storm-prediction system developed at the University of Oklahoma, which went from research to market in less than ten years, is an example of the commercial applications we might expect to see grow out of Abilene and Internet2. One of Internet2's stated goals is "to accelerate the diffusion of advanced Internet technology, in particular into the commercial sector...to sustain U.S. leadership in internetworking technology." We've only just begun to see that happen.

            In the meantime, other benefits, less tangible but just as real, are a few rails down the track. Some research, such as in the field of IPv6 studies, includes ongoing efforts of both business R&D departments and academic researchers. "Applications like the Visible Human and Digital Mammography Archive are probably most likely to come out of Internet2," Walker says. Hardware and software developments are likely to come as a result of commercial developers working in tandem with academic researchers.

All Aboard Abilene?

As it creeps out of a purely academic environment, you may find access to the new Internet at a library or museum near you within the next two or three years. That Abilene would expand into the public was the expectation from the beginning, Wood says. "The idea is to provide a high-performance window to where the Internet will be. If everything goes as planned, you and I will use it in the next five to seven years," he says.

            "The academic community proved that you can do this on a really large scale," Wood says, and new applications and middleware are already in development. In September 2001, NSF announced a $12 million initiative to develop middleware to allow other to access Intenet2.

            Anecdotal evidence indicates that commercial Internet users are ready. Walker says that as soon as his neighbors find out that he works for Qwest, the question he always gets is "When are we going to get DSL?" Although the last mile of cable going into most homes and offices hasn't changed much, Wood points out that "the backbone equipment on the commercial Internet is constantly being upgraded." So, the answer to "When will we all get onboard Abilene?" may be, sooner than anticipated. The most popular guess is that we'll be climbing onboard in the next three to five years. But, when that happens, academia may need an even newer, better, faster Internet.

            "Internet2 itself is always looking to stay three to five years ahead of the commercial Internet," says Wood. To that end, some technical groups are already calling for research beyond broadband. But, that's another story.