Time:10:00 , 3-3-2011
Place: Room 440, 4th Floor, ICT, CAS 
Speaker: Dr. Zhe Zhang IBM T. J. Watson Research Center

Abstract: 
The Internet-of-Things paradigm envisions deploying advanced sensing technologies in every corner of people’s life, and utilizing the collected information to enable intelligent decision making. An essential substrate underlying this vision is a middleware layer that is capable of analyzing massive and continuous data streams and delivering the processed information to end applications. Stream processing represents an important form of such in-network messages processing. In existing distributed stream processing systems, high availability (HA) against unexpected failures are handled by either active standby or passive standby. Our study provides an in-depth analysis of the tradeoff between active and passive standby approaches under practical settings. We find that passive standby presents a different tradeoff from active standby: longer recovery time, but 90% less overhead. We further propose a novel hybrid HA method that switches between active and passive standby modes depending on the occurrence of failure events. It achieves low overhead during normal conditions and fast recovery upon transient or permanent failure events. Experiment results show that the hybrid method can reduce two-thirds of the recovery time compared to passive standby and 80% message overhead compared to active standby, providing applications with uninterrupted processing without paying a high premium.

Bio:
Dr. Zhe Zhang is currently a Research Staff Member in the Messaging and Event Systems Department at the IBM T. J. Watson Research Center. He is also an Adjunct Assistant Professor in the Department of Computer Science at the North Carolina State University. His research has been focused on the field of distributed computing systems, with specific interests in data storage and I/O, system fault tolerance, and the application of operations research techniques in computer systems. He received his Ph.D. and M.S. degrees from the North Carolina State University in 2009 and 2006, and his B.E. degree in computer science from the University of Science and Technology of China in 2003.