[Colloquium] CANDIDATE INTERVIEW: Tim Kraska, UCBerkeley on February 29

[Katie Casey]

DEPARTMENT OF COMPUTER SCIENCE

UNIVERSITY OF CHICAGO

Date: Wednesday, February 29, 2012
Time: 2:30 p.m.
Place: RY 251

----------------------------------------------------------

Speaker:		Tim Kraska

From:		University of California, Berkeley

Web page:	http://www.cs.berkeley.edu/~kraska/

Title: 		Outlook for Tomorrow: Cloudy with a Chance of Consistency	

Abstract: There has been a lot of excitement about moving data-intensive applications to the cloud for economic reasons. Cloud computing promises virtually infinite scalability and high availability at low cost. We evaluated existing cloud services and found that they barely fulfill this requirement. Most importantly, current systems are limited by the way, how they achieve consistency. High consistency implies high cost per transaction and, in some situations, reduced scalability. Low consistency is cheaper but it might result in higher operational cost because of, e.g., overselling of products in a Web shop. Therefore, it becomes useful to see consistency as a resource that can be traded off against cost. However, some application data have such high requirements in regard to consistency and availability, that they should not even be impacted by a data center failure.

In this talk, I will show how inventory-holding models used in Operations Research can help with managing this consistency-cost tradeoff. I will then present Multi-Data Center Consistency (MDCC), which lowers the cost of achieving consistency even across data center boundaries while still achieving good performance. MDCC offers consistent replication across data centers in a single-message roundtrip time in the normal operation case, and outperforms existing systems, such as Google’s MegaStore or Yahoo’s PNuts, by either saving one or more cross-data center messages per transaction or providing stronger consistency guarantees. This work was done as part of a larger project around big data management. At the end of the talk, I will provide an overview of some of my other projects and give an outline for future work.

Bio: Tim Kraska is a PostDoc in the AMPLab, which is part of the Computer Science Division at UC Berkeley. Currently his research focuses on big data management in the cloud and hybrid human/machine database systems. Before joining UC Berkeley, Tim Kraska received his PhD from ETH Zurich, where he worked on transaction management and stream processing in the cloud as part of the Systems Group. He received a Swiss National Science Foundation Prospective Researcher Fellowship (2010), a DAAD Scholarship (2006), a University of Sydney Master of Information Technology Scholarship for outstanding achievement (2005), the University of Sydney Siemens Prize (2005), and a VLDB best demo award  (2011).

Host: Anne Rogers

Refreshments will be served in Ryerson 255 following the talk at 3:30.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.cs.uchicago.edu/pipermail/colloquium/attachments/20120220/f85cb823/attachment.htm