Output details
11 - Computer Science and Informatics
University of Warwick
Continuous sampling from distributed streams
<12> An invited extension of an ACM PODS paper, the leading conference in theory of databases, in the flagship journal of the ACM. Presents novel communication-efficient protocols for continuously maintaining a sample from k distributed streams – a fundamental problem in the management of large distributed data sets. Recognised by Woodruff (IBM Research) as “initiating the study of sampling in distributed streams”. Journal version has 450 downloads. The research has impacted on work on randomised distributed algorithms (Huang, HKUST), streaming data warehouses (De Rougemont, CNRS Paris), and approximate maximum matching (Huang, Microsoft Research). The research is now protected by US Patent 8,458,326.




