"Automating Database Design for Streaming Workloads with Shared Mutable State"
Wednesday, May 24, 2017 at 12:00 Noon
Room 368 (CIT 3rd Floor)
In the past, streaming data management systems (SDMSs) have eschewed transactional management of shared mutable state in favor of low-latency processing of high-velocity data. Streaming workloads involving ACID state management have required custom solutions, usually involving a combination of database systems. This proposal suggests S-Store as a solution: a novel transactional streaming system with best-in-class performance on hybrid streaming and OLTP workloads. To achieve optimal performance, distribution in a cluster is necessary, which, in turn, creates unique challenges, as well as opportunities. The proposed research seeks to define design best-practices for distributed data management in a transactional streaming engine, as well as create an offline automatic database designer that can distribute state and processing according to those principles.
Host: Professor Stan Zdonik