Electrical Engineering and Computer Science

CSE grad students win best student paper at OSDI 12 symposium

Their paper addresses the challenge of troubleshooting the performance of production software.
Mona Attariyan Enlarge
Mona Attariyan
Michael Chow Enlarge
Michael Chow

Recent CSE PhD Mona Attariyan and PhD candidate Michael Chow, along with their advisor, Professor Jason Flinn, have won the Jay Lepreau Best Student Paper Award at the 10th USENIX Symposium on Operating Systems Design and Implementation (OSDI ’12), which took place October 8 – 10 in Hollywood, CA.

The paper, entitled “X-ray: Automating Root-Cause Diagnosis of Performance Anomalies in Production Software,” addresses the challenge of troubleshooting the performance of production software. Most existing tools, such as profiling, tracing, and logging systems, reveal what events occurred during performance anomalies. However, users of such tools must infer why these events occurred; e.g., that their execution was due to a root cause such as a specific input request or configuration setting. Such inference often requires source code and detailed application knowledge that is beyond system administrators and end users.

The paper introduces performance summarization, a technique for automatically diagnosing the root causes of performance problems. Performance summarization instruments binaries as applications execute. It first attributes performance costs to each basic block. It then uses dynamic information flow tracking to estimate the likelihood that a block was executed due to each potential root cause. Finally, it summarizes the overall cost of each potential root cause by summing the per-block cost multiplied by the cause-specific likelihood over all basic blocks. Performance summarization can also be performed differentially to explain performance differences between two similar activities.

The researchers propose X-ray as a tool that implements performance summarization. Their results show that X-ray accurately diagnoses 17 performance issues in Apache, lighttpd, Postfix, and PostgreSQL, while adding 2.3% average runtime overhead.

Mona Attariyan received her PhD in Computer Science and Engineering in May 1012 and now works at Google Seattle. Her research interests broadly include software systems with an emphasis on operating systems, with a specific focus on software reliability.

Michael Chow is second year graduate student interested in software systems research, working with Prof. Flinn in the Pervasive Computing Research Group.

Prof. Jason Flinn’s research interests include operating systems, mobile computing, storage, and distributed systems. He is currently interested in creating software systems that allow concurrent programs to execute more reliably on multicore computers, as well as tools that help developers and administrators troubleshoot software problems such as misconfigurations. He is also interested in enabling demanding applications to run on small, mobile computers, as well as consumer electronic devices such as smartphones.

Algorithms, Languages & Databases; Graduate students; Research News; Student News