Database System Design for Cloud Computing
Thursday, December 12, 2024, 11 a.m.
Bahen Centre for Information Technology, BA 3200
This lecture is open to the public. No registration is required, but space is limited.
Abstract:
Much of the research in data management involves redesigning system components to use cloud computing platforms. Cloud computing has many characteristics that differ from on-premise servers that were the target of past data management systems. For example, it traditionally disaggregates storage services from compute servers. The latest cloud platforms also disaggregate memory from compute servers, and they offer different types of auxiliary processors, such as GPUs and SmartNICs. They exhibit performance improvements, such as faster networking, but also performance degradations, such as higher storage latency. These changes demand reconsideration of past designs of data management components. In this talk, I’ll summarize these trends and show their impact on data management systems through research projects I’ve pursued in recent years, such as distributed transaction processing, remote caching, and storage offloading.
Bio:
Philip A. Bernstein is a Distinguished Scientist at Microsoft Research and Affiliate Professor at University of Washington. He was previously a product architect at Digital Equipment Corporation, a professor at Harvard University, and VP Software at Sequoia Systems. He has co-authored over 200 papers on database systems and two books on transaction processing. He is a Fellow of the ACM and AAAS, a winner of the E.F. Codd SIGMOD Innovations Award, and a member of the U.S. National Academy of Engineering. He received a B.S. from Cornell and MSc and PhD from University of Toronto.