Website for DSPA Spring 2021
This project is maintained by vasia
When: Tuesday, April 20, 9:30am ET
Modern applications in many domains now operate on high-speed streaming graphs that continuously evolve at high rates. Efficient querying of these streaming graphs is a crucial task for applications that monitor complex patterns and relationships. This talk will present our recent work on continuous query evaluation over streaming graphs within the context of the s-Graffito project. First, I will outline the common characteristics of applications that query and process streaming graphs and describe the requirements for a general-purpose streaming graph query processing framework. Next, I will introduce a streaming graph model and algebra that describes the precise semantics of persistent graph queries. Our Streaming Graph Algebra (SGA) constitutes the logical foundation for evaluation, planning and optimization of streaming graph queries. Finally, I will present our prototype implementation of a streaming graph query processor based on the proposed SGA. In particular, I will describe efficient physical implementations of SGA operators that utilize the temporal patterns of sliding window movements over streaming graphs. Our prototype implementation compiles streaming graph queries into dataflow computations consisting of SGA operators and demonstrates the feasibility and the performance of our algebraic approach for persistent query processing over streaming graphs.
Anıl Paçacı is a PhD candidate in the Cheriton School of Computer Science at the University of Waterloo. His research interests are in scalable systems for management of streaming data, with a primary focus on graph-structured data. He is currently working on algorithms and system architectures for persistent query execution on streaming graphs. Previously, he obtained his MSc in Computer Engineering from Middle East Technical University where he worked on systems to utilize observational healthcare data in drug surveillance studies.