Raman, Vaidya win Best Paper Award
Vijay Raman, a graduate student in electrical and computer engineering, and CSL researcher Nitin H. Vaidya, an ECE professor, are trying to find a new solution to an old engineering problem: How to get the most out of what you have.
While the problem of maximizing finite resources might be an old one, Raman’ s solution could usher in a new era of wireless communication. Raman studies multichannel wireless networks and ways to more efficiently and quickly transmit data.
For his paper titled “SHORT: A Static-Hybrid Approach for Routing Real Time Applications over Multichannel, Multihop Wireless Network,” Raman was awarded the Best Paper Award at the Eighth International Conference on Wired/Wireless Internet Communications in Lulea, Sweden. The paper examined ways to efficiently transmit real-time data—such as voice—over a multichannel wireless network. In a typical wireless network, different routers perform different tasks based on user needs. One router may be used to communicate with another to stream the audio and video of a Skype call, while a router in the vicinity sends text data in the form of an e-mail to another.
However, if all of the tasks take place on the same channel, there will be interference between the different types of data. So, in a multichannel network, each of these tasks is organized to take place on different channels. Just like radio stations play different music at different predetermined frequencies, routers switch channels to transmit different types of data to different users.
This switch can cause a delay in the transfer of data of between 5 and 10 milliseconds, and because information is often sent over long distances involving multiple routers, these millisecond delays can add up, slowing a system significantly. As a result, the audio of a Skype call over this network, for example, may sound jarred and distorted to the receiver.
“When we have a multihop route in a multichannel communication, I want to know how to minimize the channel switching delays without affecting the throughput of a file transfer, for instance. Essentially, the paper deals with a routing and channel allocation problem that tries to achieve this objective,” said Raman, who is advised by Vaidya.
In his paper, Raman detailed the experiments that he ran on a mesh network test bed of routers placed throughout the fourth floor of the Coordinated Science Laboratory. Using a protocol Raman developed on a Linux kernel, he managed to improve the average delay due to channel switching from 200 milliseconds to just three milliseconds.
“Everything in the paper is experimental. So, instead of saying, ‘This might work in theory,’ we know this actually achieves and I think that’s one of the biggest reasons my paper was selected for the award,” said Raman.
A mesh wireless network is a network made up of various nodes—whether these are a wireless router, cell tower, or iPhone—that can communicate information. Though these networks are desirable because information can reach its destination through variety of paths, they are also difficult to operate efficiently because of the delays and throughput issues detailed in Raman’s paper. But, as the ability to operate these networks becomes more efficient, thanks to efforts by researchers like Raman, the future of wireless communications may change radically.
It may one day be possible that no individual or business will need a private Internet subscription. Instead, anyone with a wireless router will be able to access the central Internet bandwidth provided by a state or municipality by routing it through the other nodes in a network. Instead of paying for an Internet cable running into your home, a city government, such as Urbana, would provide its citizens with bandwidth wirelessly and for significantly less expense.
This vision of the future is still far off into the future. Numerous complex problems would exist in trying to engineer such a system, most notably organizing a giant collection of private routers so that they do not interfere with one another. This coordination of nodes is a problem that Raman and others in the field will be tackling for years to come.
“There’s a lot of research to do in the future. We have a scheduling problem to deal with and a spectrum allocation problem. How many channels do we need and what communications take priority and how do we divide them by channel,” said Raman. “There’s a lot of theory, but we need to develop them in reality.”