Gig.U Project Aims for an Ultrafast US Internet
The US lags behind numerous other countries in the deployment and affordability of high-speed Internet services.
The recent State of the Internet report (www.akamai.com/stateoftheinternet, registration required to access document) by Akamai, an Internet content distribution network, places the US in 14th place globally in terms of average broadband speed, behind countries such as South Korea, Japan, and Romania.
In response, a new initiative called the University Community Next Generation Innovation Project (www.gig-u.org) — commonly known as Gig.U — is promoting the adoption of gigabit-per-second service to US homes and businesses.
Gig.U is focused on enabling participants to share information about the best and most economical practices for building wide-area gigabit networks.
The group has already attracted 37 sets of partners — each set includes a university on one hand and local ISPs, government planning managers, and businesses on the other — willing to deploy ultrafast systems.
Participants will be responsible for designing, financing, and, in the case of ISPs, constructing the networks.
While US high-speed systems such as Internet2 and National LambdaRail address backbone networking, Gig.U aims to reduce the cost and complexity of networks running from the ISP to the customer.
The project will focus on developing high-speed networks to serve neighborhoods near college campuses for the time being, with an eye toward doing so in other areas in a few years, once initial problems have been identified and addressed.
The new services could improve the performance of already-popular applications such as streaming high-definition video, online games, and videoconferencing, which sometimes suffer from latency on today's networks.
However, the real benefit would come from the new applications that the gigabit systems would enable, said Gig.U executive director and founder Blair Levin.
These applications could include high-speed video rendering and analysis of information from magnetic resonance imaging and other data-intensive medical scans.
The Gig.U program could help with this because in the past, innovative broadband applications have developed only in places where bandwidth wasn't a barrier, noted program director Ellen Satterwhite.
Despite the promise that Gig.U's goal of widespread high-speed-network development offers, the project faces several important challenges.
Setting the Stage for Gig.U
Levin, an attorney and telecommunications expert who used to work for the US Federal Communications Commission, started Gig.U, inspired by the National Broadband Plan for improving US broadband Internet service.
The FCC introduced the national plan, which Levin helped develop, in 2010. One of its goals was providing 100 million American homes with access to 100 Mbps connections by 2020.
According to Akamai director of market intelligence David Belson, today's high-speed networks worldwide operate at a peak of from 20 to, in rare cases, 100 Mbps.
Akamai's State of the Internet report found the highest average broadband speed worldwide was 21.8 Mbps in Daegu South Korea, followed by several other South Korean and Japanese cities. The US city with the highest average broadband speed was Boston, with only 8 Mbps.
Belson said a key issue in addition to speed is affordability. For ultrafast service, consumers might pay $40 per month in South Korea but $400 in the US. In some cases, governments offer incentives that enable local carriers to offer lower rates.
Levin noted, "We found that America does not have a plan for an upgrade. We do not have realistic expectations that America will have the best networks in the world. We do not have a national provider to build a better network than the best currently available network."
Globally, said Belson, "The places with the highest penetration of high-speed connectivity in many cases are being driven by high population density, government support, and in some cases, a national initiative."
US networks have kept pace with popular consumer applications' bandwidth requirements, so service providers have seen no need to plan for and invest in significant upgrades that current programs don't need, Levin explained.
He contended that instead of waiting for new applications that justify creating high-speed networks, the US needs to create the technology for reasonably priced networks that will encourage the development of new programs.
The Google Community Fiber Initiative (https://www.google.com/fiber), announced in 2010, has deployed a small demonstration gigabit network in homes near Stanford University and hopes to build a prototype system near Kansas City, Missouri in the near future.
ISP Sonic.net has deployed a small scale consumer gigabit network to homes in Sebastopol, California.
Case Western Reserve University has rolled out a gigabit network to about 100 homes and businesses in a Cleveland neighborhood.
Start-up Surgical Theater is using the Case Western network in development of its highly demanding surgical simulator.
Gig.U: Up Close and Personal
Gig.U members include Case Western, Columbia University, Duke University, Indiana University, Michigan State University, North Carolina State University, the University of Florida, and the University of Maryland.
Levin said he decided to work with universities because they are already connected to high-speed research networks such as Internet2 and National LambdaRail, which could provide the backbone for gigabit last-mile networks.
In addition, he noted, universities have many heavy Internet users.
Gig.U is focused on both network technology and applications that could use the technology.
The project intends to find ways for participants to leverage commercial technologies, low-cost construction techniques, and best practices to develop an affordable broadband-network infrastructure.
Gig.U members and participating ISPs would fund the development of the networks.
The organization held a request for information last year that drew ideas from 24 service providers and 14 network-equipment vendors.
One of the ISPs, GWI, has committed to building a gigabit network near the University of Maine.
Gigabit Squared, a start-up supported in part by equipment and software vendors, promised to invest up to $200 million to develop as many as six for-profit networks in partnership with Gig.U members.
According to Gig.U's Satterwhite, the project will remain technology neutral so that each participant can use the approaches they deem most suitable.
Gig.U expects participants to use a variety of existing technologies, including fiber optics.
Gigabit Squared president Mark Ansbury said his company is exploring wireless technologies such as free-space lasers, which transmit data via light propagating through the air; and WiMax, a radio-based approach.
He noted that Gigabit Squared is also studying ways to reduce the cost of connecting users.
For example, with microtrenching, providers could accurately dig narrower, shallower trenches for optical fiber, thereby reducing costs, construction time, cleanup, and the risk of hitting utility lines.
Ansbury said Gigabit Squared is also investigating robotic equipment, as well as precut and preconnectorized cables, to decrease installation expenses.
Up and Away?
The Gig.U project faces a number of technological and operational challenges.
Few applications today require or interest users in ultrafast networks.
The new networks would entail significant installation costs and construction-related disruption in surrounding neighborhoods.
In fact, Gig.U's Levin said, a key challenge will be reducing network-development costs enough to make the benefits worthwhile.
Akamai's Belson noted, "The million-dollar question is: Once you have an enormous pipe to the home, then what? It is a great proof of concept, but we have not figured out how to fill a gigabit per second."
However, Ansbury said, higher speeds would make it possible to move many services and applications to the cloud.
For example, he noted, various home devices gather data for security, entertainment, and healthcare applications. The systems must process the large amounts of information locally, burdening users' computers in the process, to reduce demand on today's limited networks. Gigabit networks could change this.
Levin stated that applications like genetic sequencing, robotics, or high-definition video require the additional bandwidth such networks would provide but that the most exciting programs probably haven't been developed yet.
He predicted that the resulting higher bandwidth would lead to new applications and new ways of using the Internet.
"It took 40 years for 50 percent of American factories to utilize distributed electricity," Levin said. "I don't think it will take that long to develop the applications that 50 percent of users would need a gigabit network for."