Data Centric Networking and Systems (2012-2013 Lent Term)
|
|
Guest LectureTitle: "Drinking from the Fire Hose: The Rise of Scalable Stream Processing Systems " (slides) Speaker: Peter Pietzuch, Large-Scale Distributed System Group, Department of Computing, Imperial College London Abstract: We witness an increasing number of domains that rely on just-in-time processing of large volumes of data. Financial trading, web analytics, fraud detection and sensing applications all require the continuous processing of data streams. This processing model is not well-suited for classical database management systems (DBMS), which have storage-centric designs and assume that data changes less frequently than queries. Research on distributed data stream management systems (DSMS) has questioned these assumptions and provided models and architectures for the efficient execution of continuous queries over data streams. In this talk, I will give an overview of the field of DSMSs, describing some of the most influential work in the past, and discuss open challenges that DSMSs face when deployed across a large number of servers in a cloud setting. |