University of California, San Diego (UCSD), computer scientists have developed an automated, enterprise-scale Wi-Fi troubleshooting system. UCSD computer science professor Stefan Savage said people expect Wi-Fi to work, but have a general understanding that Wi-Fi systems can be inconsistent.

“If you have a wireless problem in our building, our system automatically analyzes the behavior of your connection—each wireless protocol, each wired network service and the many interactions between them,” Savage said.

Manually diagnosing wireless access networks requires a massive amount of data, expertise and time. And wireless networks contain complexities that wired networks avoid, such as problems in the shared spectrum, user mobility and authentication management. Additionally, the interaction between wired and wireless networks also can be the source of problems.

“Many problems are transient—they’re gone before you can even get an admin to look at them—and the number of possible reasons is huge,” said Yu-Chung Cheng, UCSD computer science Ph.D. student.

 In a paper on the system presented at ACM’s SIGCOMM 2007 conference, the researchers outlined a set of modeling techniques for automatically characterizing the source of Wi-Fi problems, with a specific emphasis on data transfer delays unique to 802.11 networks. The system is used in the UCSD Computer Science building, and all wireless help desk issues are directed to the automated system, which at the time of this publication had been running for about nine months.

“In the future, I think that enterprise wireless networks will have sophisticated diagnostics and repair capabilities built in,” Savage said.  EC