Taming Graph Dynamics at Scale - Felix Cuadrado

Real-world graphs are dynamic, as they are created from messages, events, infrastructure and interactions occurring at certain points in time. However, most large-scale graph analysis frameworks focus on processing static graphs, or 'snapshots'. In this talk I show how Pregel's computation model can be adapted to dynamic graph processing, discuss the performance effect of graph partitions for dynamic graphs, and present a new adaptive partitioning technique designed to work on graph changes. Dr Félix Cuadrado is a Lecturer (Assistant Professor) in the School of Electronic Engineering and Computer Science, QMUL. He is the Program Coordinator of QMUL's MSc in Big Data Science. Felix has spent the last 8 years researching and building distributed systems, and is interested in understanding distributed systems performance, and engineering management techniques for them. http://datasciencelondon.org

Comments