Browse by author
Lookup NU author(s): Visalakshmi Suresh, Dr Paul EzhilchelvanORCiD, Professor Paul WatsonORCiD
Event processing involves continuous evaluation of queries over streams of events.Response-time optimization is traditionally done over a fixed set of nodes using avariety of techniques. The emergence of cloud computing makes it easy to acquireand release of computing nodes as required. Leveraging this facility, we propose anovel, queuing-theory based approach for meeting specified response-time targetsagainst fluctuating event arrival rates by dynamically and adaptively drawing anadequate amount of computing resources from a cloud platform. In the proposedapproach, the entire processing engine of a distinct query is used as an atomic unit foroptimization and reconfiguration. Several such units hosted on a node are modelled asa multiple class M/G/1 system. These aspects not only eliminate the need forintrusive, low-level performance measurements during run-time, but also offerportability and scalability. The efficacy of the approach is examined through cloudbasedevent- processing experiments where dynamism and adaptation are shown to beachievable.
Author(s): Suresh V; Ezhilchelvan P; Watson P
Publication type: Report
Publication status: Published
Series Title: School of Computing Science Technical Report Series
Year: 2012
Pages: 16
Print publication date: 01/04/2012
Source Publication Date: April 2012
Report Number: 1324
Institution: School of Computing Science, University of Newcastle upon Tyne
Place Published: Newcastle upon Tyne
URL: http://www.cs.ncl.ac.uk/publications/trs/papers/1324.pdf