Just because the commercial world has taken over the Internet, academic and (US) government researchers had to go off and build their own private network and dub it Internet2. Here we sit, with our noses pressed to the window, while they tap into what is now the world's fastest wires for their esoteric investigations into things like prehistoric weather patterns. It's as if all the commercial activity on the Internet has polluted the network that academics and the government originally developed for their own use.
There's just one problem. It's not true. Contrary to popular belief, the so-called Internet2 project isn't a separate network of trenches and fibre meant to keep out the commercial masses. Nor will there ever be two Internets that criss-cross the globe. Instead, think of Internet2 as a giant test bed, where more than 200 universities, government researchers and high-tech companies participate in research, with their efforts coordinated by the University Corporation for Advanced Internet Development (UCAID, www.ucaid.org) in Washington, DC.
UCAID's goal: to develop fundamentally different technologies that can be dropped into the current Internet protocol infrastructure as they become ready for prime time. Internet2 participants emphasise fundamentally different; research initiatives go beyond boosted bandwidth to include middleware, protocols and, most important, new Internet applications that duplicate and even improve upon face-to-face interactions.
Such applications, already in prototype, bear little resemblance to what the world has seen: three-dimensional CAT scans that show a brain in mid-thought; live feeds from upper-atmosphere observatories that let astronomy students look and listen as leading experts discuss what they see; virtual-reality-residing design studios that dissolve barriers of time and place, allowing better collaboration among global teams.
If unleashed over public networks, these bandwidth-hogging applications would bring the current Internet to its knees. Internet2 is the venue where such high-performance network applications can be tested. "This is an experimental infrastructure running experimental research-oriented traffic," says J Gary Augustson, vice provost for information technology at Pennsylvania State University in University Park, who previously chaired the steering committee that launched Internet2 in 1996. "This isn't for aggregated e-mail or putting Christmas movies online. It's a concerted effort to shorten the development life cycle of the next stage of Internet products," he says.
According to many researchers associated with Internet2, technologies being worked on now will begin to find their way into commercial products by the end of 2000. "Public networks are approaching a state of operation where they look, act and feel like private networks," says Guy Cook, vice president of Internet2 services at Qwest Communications International in Denver. Private, value-added networks are speedier and more reliable than today's public networks but are also more expensive. Beefing up the Internet with technologies that can replicate the performance of private networks will be a significant advancement in networking technology, says Cook.
"That completely changes the economics," he says, "because instead of having to construct a private network, CIOs can soon turn to the massive telecommunications companies to deliver those capabilities on the public infrastructure." Eventually, the Internet as we know it will be replaced piecemeal as new technologies and components derived from Internet2 get commercialised. The result for the world's businesses is that everything changes -- from strategic business models to business processes and staffing skills. Companies will have to rethink how they work, how they build and sell products, and how they manage their network assets. In an age of Internet time, it becomes even more critical to understand high-speed network technologies and the capabilities they will enable.
The Internet2 project is exploring a host of technologies related to the fast and reliable transmission of video and audio. Two such technologies -- multicasting and differentiated quality of service -- could be among the first to make it outside the research lab.
The notion behind multicasting is sending a single stream of video or audio across the Internet and then, at the closest common point on the network, copying it to different computers requesting that same stream. The result is vastly improved network performance because each individual computer doesn't have to query the originating server separately for data.
Then there's differentiated quality of service, or diffserve. Today Internet transmissions are "best effort" -- packets get there when they get there. That's not a problem with e-mail messages, which don't have to be delivered instantaneously. But packets in a videoconference stream need to reach their destination without even a millisecond break. Diffserve would take care of the issue by assigning levels of priority to different types of packets. Combined, the two technologies will allow the intercast of both live and on-demand, TV-quality video and audio. That may not seem like much, but it's revolutionary because it will reduce -- if not eliminate -- the confines of both time and space. And that's the overarching focus of Internet2.
"The key to this whole exercise is applications, not infrastructure," says Terence Rogers, who -- as director of the 2.4Gbps backbone known as the Abilene Project that connects Internet2 members -- oversees the fastest infrastructure component ever to link US universities. "The whole point of the infrastructure is to allow the development of experimental, multimedia applications as a way to stimulate the commercial sector to deploy more bandwidth."
The rationale, Rogers says, is to crack the chicken-egg conundrum. Today commercially available broadband capacity is expensive, primarily because few applications demand it. But no one's writing bandwidth-consuming applications, since there's no way to run them. "We hope that by encouraging people to create experimental applications, we'll stimulate the commercial sector to deploy more bandwidth," says Rogers.
The applications developed for Internet2 themselves seem lifted out of Star Trek. Consider, for example, the virtual reality "cave" developed at the University of Illinois. Everyone wearing the requisite VR goggles, no matter where they are, sees the same 3-D images projected on a wall. Onscreen (or is it onwall?) adjustments and changes are seen by everyone.
Or imagine a videoconference in which the images of participants are captured by a 180-degree surround camera and then projected thousands of miles away onto a wall. Life-size. With directional voice. "It's like being there," says John Patrick, vice president of Internet technology at IBM in Somers, New York. "I think of these applications as natural-net rather than Internet."
Not surprisingly, nearly all of the experimental applications now in prototype have an educational focus such as distance learning. But the combination of high-bandwidth, full-screen video and CD-quality audio creates extraordinary possibilities for business use too: visualise stock market data in 3-D. Test drive cars before they're built. Reach out to customers with full-length, high-quality video presentations at their desktops.
This isn't so much as thinking outside the box as it is demolishing the box. The ramifications for commercial use could be far-reaching. "By 2000, we will see some businesses using these capabilities as a competitive edge to gain market share and customers," predicts Joe Mambretti, director of the International Centre for Advanced Internet Research (iCAIR, www.icair.org) at Northwestern University in Evanston, Illinois. (iCAIR works jointly with UCAID on Internet2 research projects.) "These technologies will change whole business strategies in how companies define themselves."
Of course, for all of the blue-sky concepts to come to fruition, honest-to-God products that incorporate multicasting and diffserve technologies must hit the market. That means products from the likes of Ameritech, AT&T, Cisco Systems, IBM, Lucent Technologies, MCI WorldCom, Nortel Networks and Qwest. These represent many of the heaviest hitters in the Internet telecommunications arena, and all are corporate Internet2 members. As such, each has pledged more than $US30 million to Internet2 research over the life of the project. Each undoubtedly expects big returns.
Take Qwest, for example, which donated 11,000 miles [17,600 kilometres] of 2.4Gbps (OC-48) capacity on its nationwide network. Value of the donation: $US500 million. But Qwest didn't just hand over bandwidth and step back. The company is working shoulder to shoulder with all of the other members to learn firsthand what and how things break at super-fast speeds. "Qwest doesn't do a lot of direct R&D what better place to get advanced applications from?" says Qwest's Cook. A big plus of Qwest's participation in Internet2, says Cook, is observing how sufficiently different networking equipment such as routers and protocols work together. Cook says Qwest will edge out competing vendors by directly applying its Internet2 gleanings to its existing infrastructure. "We plan to be very aggressive in getting AT&T customers to move from their expensive traffic transmissions to us," Cook says. (Of course, fellow Internet2 member AT&T has its own ideas about keeping and gaining customers.) Even companies that do a lot of direct R&D are involved with Internet2. At IBM, the mantra for the decade is "e-business". And why not? IBM has generated nearly 25 per cent of its 1998 revenues of nearly $US82 billion from hardware, software and services related to the company's e-business offerings. With that kind of money at stake, IBM has good reason to participate in just about every major initiative of the Internet2 project. IBM is learning how to manage and monitor ultra-fast networks for authentication, security, accounting procedures, performance monitoring and network management. Members of IBM's Global Services consulting practice are currently immersed in Internet2 research projects, says Patrick. He expects Big Blue to offer equipment using Internet2 technologies within the next 12 months. Among the specific Internet2 initiatives IBM is touting are high-speed, videoconferencing and distributed network storage.
Is there anything CIOs can do now to prepare? Yes and no. No, in that nothing is yet available to buy or contract for. Yes, in that now is the time to plan for the coming Internet2 revolution.
"The level of competency an IT organisation needs in the network arena is the biggest issue to deal with," says Erv Blythe, vice president of information systems at Virginia Polytechnic Institute and State University in Blacksburg. "The network becomes the vital infrastructure, and that means there will be a premium on hiring people with networking skills, maintaining their knowledge and retaining them."
Blythe should know. As an IS executive of an Internet2 member, he's among the first to see how IT organisations must change and adapt. Among his discoveries: IT organisations must embrace the new world of directory-enabled applications to manage and coordinate the new Internet. "A few years ago, CIOs spent their time with the database management system," says Blythe. "Now, they have to turn their attention to directory management for assigning levels to users and for establishing policies in how people and data move through the network."
And finally, Blythe advises, all assumptions about the networking industry are off. Think you know who the leaders in networking and telecommunications are? They might not be the dominant players eight months from now. "We'll see some awfully big players drop out of the Internet communications arena, and some unknowns take their place," says Blythe. "So you have to keep your options open -- don't think in terms of many-year partnerships. And always keep the number-two and number-three guys on your radar."
Join the CIO Australia group on LinkedIn. The group is open to CIOs, IT Directors, COOs, CTOs and senior IT managers.