Performance Modeling and Anomaly Management for Complex Systems

Dynamic-content online services are comprised of multiple interacting components and data partitions distributed across server clusters. Understanding the performance of these services is crucial for efficient system management and capacity planning. This project investigates profile-driven performance models for multi-component, data-intensive online services. Research has focused on the characterization of inter-component communications as well as concurrent data accesses to storage devices. Built on the performance modeling, research also studied the identification of performance bug manifestations and comprehensive depiction of performance anomaly manifestations over a large space of system configurations and runtime conditions. In particular, our work has resulted in the identification of previously unknown I/O performance bugs in the Linux operating system.

Most Relevant Publications:

Project Members:

Chuanpeng Li, graduated with Ph.D., now at Bloomberg
Xin Li, University of Rochester
Pin Lu, graduated with M.S., now at Microsoft
Stan Park, University of Rochester
Kai Shen, University of Rochester
Christopher Stewart, graduated with Ph.D., now on the Computer Science faculty at the Ohio State University
Ming Zhong, graduated with Ph.D., now at Google

Contact:

Kai Shen (kshen@cs.rochester.edu)

Support:

This project is supported by the National Science Foundation under Grants No. 0615045 and No. 0448413.

Note:

Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.