StreamMapReduce - When Stream Processing Crosses MapReduce
Speakers: André Martin
Systems Engineering Group
Dresden University of Technology
November 3, 2015 - 4:00 p.m.
Location: Sage 5101
Hosted by: Carlos Varela (x6912)
Event Stream Processing (ESP) is a well-established approach for low-latency data processing enabling users to quickly react to relevant situations in soft real-time. Although such systems exist for already more than a decade, we recently witness a true renaissance for ESP systems that have adopted the popular MapReduce paradigm. Examples for such systems range from Apache S4, Storm to Samza. In this talk, we will discuss the properties of StreamMapReduce and present StreamMine3G, an ESP system employing the StreamMapReduce approach. We will cover the design and architecture of the system, provide real world applications examples that have been implemented on top of StreamMine3G and present approaches for fault tolerance and elasticity.
André Martin is a fifth year PhD student with a research focus on low overhead fault tolerance in Event Stream Processing systems. Prior to working on his PhD and joining the Systems Engineering group at TU Dresden, he worked at the high performance computing center (ZIH) of the TU Dresden.
Last updated: October 30, 2015