Internet2 turns 15: Has it delivered on its promise?
- — 11 April, 2011 23:28
With nearly $100 million in new funding, Internet2, the faster, better Internet reserved for research and education, has embarked on an upgrade that will boost backbone capacity to a staggering 8.8Tbps and expand services to hundreds of thousands of libraries, schools and medical centers.
TIMELINE: Internet2 milestones
Internet2 was created by 34 university research institutions in 1996, when the commercial and non-commercial branches of the Internet's evolutionary tree split off and went their separate ways. The mission of Internet2 was to provide reliable, dedicated bandwidth to support the ever-growing demands of the research and educational communities, and in doing so, to develop technologies that would advance the state of the 'commodity' Internet.
As Internet2 celebrates its 15th anniversary, there's widespread agreement that the non-profit organization has succeeded in the first part of its mission. Internet2 delivers hybrid optical and routed IP-based network services to more than 300 member organizations.
The Internet2 backbone is an Infinera-based optical network that delivers up to 400Gbps of wavelength capacity. The network is operated in partnership with Level 3 Communications. And Juniper routers are used to create multiple 10 Gigabit Ethernet links between nine core nodes around the country.
Downstream, Internet2 connects to 20 regional optical networks at speeds ranging from 10G to 30Gbps. These regional networks ultimately deliver broadband connectivity to more than 66,000 anchor sites, such as museums, performing arts centers, libraries, K-12 schools and medical centers.
Internet2 is also the link between the U.S. and the global research community. For example, U.S. scientists involved in the Large Hadron Collider particle physics research in Europe could not participate without Internet2.
"Currently, the largest users of the network are scientific, and the Large Hadron Collider (LHC) scientists' usage dwarfs most of the other disciplines," says Jason Zurawski, network software engineer at Internet2.
Each of the LHC facilities worldwide has multiple petabytes of storage and thousands of computing cores to process and analyze the data, which must be distributed via networks such as Internet2 for real-time collaboration and iterative processing of datasets that can reach 100TB, he adds.
Internet2 has a long list of other achievements and accomplishments. As Rodney Wilson, senior director of external research at Ciena, points out, "Internet2 creates temporary networking test beds, facilitates research on next-generation networking concerns, and has contributed to a number of breakthroughs and world firsts, such as the first 100 Gigabit Ethernet research network and the first uncompressed high-definition videoconference using dynamic circuit networks."
Wilson adds that Internet2 is more than just the network infrastructure. It provides its members with tools for network research, middleware used for provisioning network services, network performance, and security. Several of these tools like Shibboleth (identity and access management), perfSONAR (network performance measurement), and OSCARS (network provisioning) have been adopted by researchers worldwide and have helped address some of the most significant global advanced networking concerns and issues.
Gregg Kreizman, research director at Gartner, is impressed with the progress Internet2 has made with identity and federation. "I find Internet2's work in this area remarkable because of its technology, the resultant benefits of its use (particularly resource sharing and collaboration), the community, identity governance, and topic education."
He adds, "I point clients in all industries to Internet2/Shibboleth as a solid example of open-source federation technology they can use as an alternative to commercial offerings. With any new or evolving technology and with any large communities, there are pitfalls along the way, but I think Internet2/Shibboleth and InCommon (the federation developing a common framework for access control) have emerged as very positive influences for identity and access management."
Scott Crawford, managing research director at Enterprise Management Associates, describes Internet2's work as "pioneering." He says, "Shibboleth was one of the first practical implementations of identity federation."
And there's no shortage of Internet2 end users who sing the praises of the super-fast network for enabling applications like telemedicine and distance learning.
Dr. Dale C. Alverson, a professor of pediatrics and medical director at the Center for Telehealth and Cybermedicine Research at the University of New Mexico, says, "We have used Internet2 for virtual reality simulations where participants enter the virtual world simultaneously in Perth, Australia, and Albuquerque to treat a virtual patient as a team. In addition, we have trained a resident in Hawaii with a standardized patient at UNM and evaluators commenting from different locations to provide monitoring and feedback.
Helen Smith, director of technology & media production at the Eastman School of Music in Rochester, N.Y, adds, "We use Internet2 to bring other cultures into Rochester. For example, world music from Indonesia and Japan can be shared by performers in those countries, as can early music on authentic instruments from England and Germany, and near-simultaneous performances and new commissions involving Internet2 technology are possible."
Marla Davenport, director of learning and technology at an education technology consortium in Minnesota, says, "We have used Internet2 for many special student projects such as a collaborative concert with a school in Taiwan, projects on the West Nile Virus, NASA virtual field trips, iLab for science experiments, heart valve and knee replacement surgeries, and many more."
In one case, students in the Osseo and Cambridge-Isanti, Minn., High Schools watched heart surgery live along with a class from Rhode Island. The students were briefed by the surgeon who narrated the process as students watched via a head-cam attached to the surgeon. Davenport adds, "They were engaged for the entire two hours of the surgery and left with new information and perceptions about careers in the medical field."
But has Internet2 delivered on its promise to the commercial Internet? "It depends on what one thinks was Internet2's promise," says Scott Bradner, technology security officer at Harvard University. "It has provided very good connectivity for higher education research, but it has not resulted in a lot of useful network research. A number of the original network-related research projects made good progress resulting in a rather large number of research papers and quite a few advanced degrees. But some of the research was overtaken by events."
He points out that QoS research was an early Internet2 effort but it turned out that advances in the speed of the Internet rendered QoS technologies that were being worked on unnecessary. "That said, a lot of technical and application research is still going on," Bradner adds.
Tom Nolle, president of CIMI Corp., comes at the question from a different perspective. "In my view, Internet2 had two missions: a technical one and a market one. The technical objective was to explore the boundaries of Internet technology, particularly for higher speeds and more types of services (including circuit-mode connections), and the business objective was to prepare the Internet for the broadband age. I think it met the former objectives, but failed in the latter."
According to Nolle, Internet2 is relevant as a technology test-bed, but a non-profit doesn't operate the same way as a commercial enterprise and thus can't serve as a model for one. Therefore, it's been unable to contribute, in any meaningful sense, to the non-technical issues, such as public policy, or business models of the service providers, he argues.
"The bottom line," Nolle says, "is Internet2 a science project? If it doesn't have any relevance to the issues of the [commercial] Internet as it is now, then what good is it really going to do? My view is that the success or failure of any publicly-funded project is whether it benefits the public, and I suspect that most everyone in the general population would say that if Internet2 isn't going to fix problems with the [commercial] Internet, it's not helping them."
Others take exception with Nolle's viewpoint. Debbie Montano, chief architect of government, education & medical at Juniper, argues that, "While Internet2 does not operate as a 'for-profit' Internet service provider, it does; nonetheless, contribute significantly to the dialog and thought leadership on 'Layers 8 and 9' issues — particularly the financial and business considerations facing industry and government in continuing to expand the [commercial] Internet's capabilities."
"In my view," she continues, "it's extremely helpful to have an organization such as Internet2 (including its members and affiliated state/regional networking organizations) who can articulate a long-term vision for the Internet, for expanded broadband deployment, and for IPv6 adoption without being focused solely on quarterly results. The economic health of the United States and the continued vitality of the commercial Internet depend on the willingness to innovate and the thoughtful consideration of national and international networking issues at all layers, which Internet2 embodies so well."
Lauren Rotman, director of communications at Internet2, adds, "Every year, a million students graduate from universities where they have used Internet2's 100Gbps network to collaborate with faculty, colleagues, and other students. They have used applications such as HD-quality videoconferencing, video repositories, and other real-time collaboration tools on a regular basis. When these students enter the workforce, they will expect these same networking capabilities. Their opinions and recommendations will be the primary, driving force that demands upgrades to corporate networks and the commercial Internet for both applications and higher bandwidth capabilities. These are the voices of our future, the voices that will effect change."
And while there is room for a healthy debate on this issue, it seems clear that Internet2's members are satisfied with its performance and continue to support it. As Bradner says, "The member universities that pay to connect to Internet2 must think it's worth the money, which is a real statement since the Internet2 costs are in addition to the commercial Internet costs."
And Internet2 bandwidth fees are not inconsequential. For example, a research institution that wanted a direct 1Gbps connection would pay $250,000 a year. A regional network that wanted a 10Gbps connection, would pay $480,000 a year.
The next 15 years
With an infusion of nearly $100 million, Internet2 is poised to continue serving the educational and research communities for a long time. The U.S. government recently awarded Internet2 $62 million in stimulus funds to upgrade the network and to make it more broadly available.
Specifically, the National Telecommunications and Information Administration (NTIA) awarded Internet2 and partners Indiana University, the Northern Tier Network Consortium, Ciena, Cisco, Infinera, and Juniper Networks funding through its Broadband Technology Opportunities Program (BTOP).
Internet2 and its partners are investing another $33 million to complete the project, says Rob Vietzke, executive director of network services at Internet2.
The upgraded network will support up to 8.8 terabits per second of wavelength capacity at its completion and potentially four to 10 times that during its lifespan. The optical system will be provisioned on Internet2-owned dark fiber on a 16,000-mile national footprint and features 50 points-of-presence with optical interconnection at most intersections. Currently, the network transports 10, 40 and 100 Gigabit optical wavelengths, Vietzke explains.
"It's an incredibly robust optical platform. It can, for example, carry up to 880 10-Gigabit paths between Albuquerque and Denver, while simultaneously carrying another 880 10-Gigabit paths from Albuquerque to El Paso," says Chris Robb, director of operations and engineering at Internet2. "One of the most exciting aspects of the optical system is its capability to support newer technologies as they are made available. We're hopeful that we'll see individual wavelengths at 1 terabit per second in the next five years."
In addition, the upgraded network, based on Ciena's ActiveFlex 6500 platform, will have 10 core nodes with 100 Gigabit Ethernet interconnects between those nodes. Backbone nodes will be interconnected with 100G Ethernet framed circuits instead of bundles of 10G. Edge connections will eventually move to 40G and 100G out to the regional and state networks, which connect member institutions to the Internet2 backbone, adds Robb.
"Network construction has already begun and we anticipate the first coast-to-coast links to be installed and operational by this summer," says Vietzke. "We have already begun upgrading core routers and preparing for the new 100Gbps optical transport system. The entire network is scheduled to be completed by June 2013; however, we already expect to accelerate that deployment in response to partner and user demand."
The upgrade will allow Internet2 to expand its footprint to an estimated 200,000 schools, hospitals, libraries, community colleges and other anchor institutions.
Sartain is the author of "Data Networks 101" and a freelance journalist from Salt Lake City, Utah. She can be reached at email@example.com.
Read more about lan and wan in Network World's LAN & WAN section.