Mitra to debug mobile-cloud applications

Sayan Mitra, an assistant professor in Electrical and Computer Engineering, has received a one-year, $108,682 grant from Samsung’s Global Research Outreach Program to look into the problem of debugging mobile-cloud applications.

Mitra, who is also an assistant professor in the Coordinated Science Laboratory (CSL), said that he and his students first noticed this problem while working on a recent project.

“Last year we started working on distributed programs for mobile phones that control robots,” Mitra said. “(The phones) instruct the robots to do something interesting, like collaboratively find an object in a space or draw a picture.”

The high-level tasks may sound straightforward, but the programming did not prove to be easy.

“We found that it’s a nightmare,” said Mitra, who is a member of the Reliable and High Performance Computing group at CSL, and is a researcher in the Information Trust Institute. “Although we had talented students writing these programs, we found lots of bugs, which came up because of message delays and concurrency … so we said, okay, how can we systematically go about finding why this thing is not working?”

That new debugging project took-off, leading to a paper for the IEEE Real-Time Systems Symposium last year, and now the grant from Samsung.

“Samsung is a industry leader in this area of mobile computing … so the fact that they found our research interesting and possibly useful was encouraging,” he said.

In general, writing software for distributed systems can be a very challenging problem, he noted, especially when looking at types of software that control physical objects, such as the case with the phones and the robots.

“Finding bugs in such programs is hard because there are uncountably many scenarios to consider,” Mitra said. “The scenarios arise from different kinds of message delays, or failures in the communication network. What we are trying to achieve … is a systematic way of finding defects.”

Mitra's approach to this problem will be to look at the actual code of the software and to collect information from the traces of the software as it is running. These two elements will be combined into what is called static dynamic analysis algorithms to find possible bugs.
______________________

Contact: Sayan Mitra, Department of Electrical and Computer Engineering, 217/333-7824.

Writer: Elise King, CSL Communications

If you have any questions about the College of Engineering, or other story ideas, contact Rick Kubetz, editor, Engineering Communications Office, University of Illinois at Urbana-Champaign, 217/244-7716.