Georgia Tech researchers want you – and your computer.
They want to figure out how to make the Internet faster and more reliable, but to do that they need to gather data from tens of thousands of personal computers around the world.
Currently, Internet performance research almost always uses data gathered from various router points along the Internet's backbone, the high-speed pipelines that keep data moving around the globe. But George Riley, professor of electrical and computer engineering at Georgia Tech, and graduate student Robby Simpson want to use data collected directly from Internet users.
To do this they've developed an open-source software application that gathers network performance statistics such as average response time, average round-trip time, connection times, download times and number of packets sent and received.
Now they just need truckloads of volunteers to download their NETI@home application, and allow the application to send network performance information from the volunteers' computers to a server at Georgia Tech, where it will be analyzed and made available to anyone else who wants to use it for their own Internet improvement projects.
Neti@home is named after the University of California at Berkeley's SETI@home project, which uses volunteer computing power to search for extraterrestrial intelligence.
SETI distributes data collected from a high-powered radio telescope in Puerto Rico to millions of computers around the world, borrowing their processing power to analyze the data for potential evidence of extraterrestrial communications or activity.
"With NETI, we're searching for network intelligence – intelligence about the way the Internet works so we can make it work better," said Simpson.
But the data that NETI will gather is far more personal than that gathered by SETI. Riley and Simpson know that people may be leery of having data on their network usage tracked and transmitted, but promise to protect volunteers' privacy.
NETI users can set privacy protections that will determine what types of data will be gathered and reported from their computers. The reports sent to Georgia Tech are also stored on the user's computer, so the user can see exactly what statistics are gathered.
"NETI is not spyware, and in no way compromises any of your private data," said Riley. "It is nearly painless from a performance point of view, and completely private if you opt to use our most restrictive privacy setting."
Simpson and Riley said they hope, however, that volunteers will not choose the most restrictive privacy setting. Less restrictive settings will allow them to gather more useful data from participating computers.
"We will use the collected data in our research to create realistic simulation models of typical Internet users' behavior," Riley said. "Better simulation models lead to better simulations, which lead to better protocol analysis, better protocols and, eventually, a better Internet."
The data collected by NETI, sans anything that might personally identify volunteers, will also be made available to other network researchers and the general public on the NETI website. As the project picks up speed – currently there are only a few dozen volunteers – they expect to make the data available in real time.
If NETI is installed on enough machines with a broad-enough distribution, the data that's being collected could even provide an early-warning detection system for worms and viruses, said Simpson.
"If we start noticing that many NETI@home users are receiving anomalous traffic, that could be an indication of the spread of an Internet worm, or some other sort of attack," Simpson said. "If the clients were distributed enough, one could even see which parts of the world are attacked first and then possibly use the data to track where the worm seems to have originated from."
Simpson also envisions using NETI data to produce a chart of the best and worse Internet service providers, in terms of performance and security.
"The beauty of collecting data such as this, data that hasn't been gathered before, is that we can't imagine all of the possibilities," said Simpson. "But one immediate way of meeting our goal of 'making the Internet better' would be to identify some of the worst network routes, and point these out to the organization that owns them. I would imagine it would be quite embarrassing for a major ISP if they were found to have the worst connections."
Systems administrators can also use NETI to monitor performance on their private networks. And since the NETI application is open-source, Simpson and Riley also hope the development community will come up with other interesting uses for the application.
Georgia Tech plans to keep the NETI project going "theoretically forever," according to Riley.
"Everyone agrees that the Internet is continually changing and growing, and data from year 2004 will likely be obsolete by 2005," Simpson said. "Besides, you can never have too much data."
