Tuesday, June 5th 2018
10:00am – HFH 1132
Committee: Divyakant Agrawal (Co-chair), Amr El Abbadi (Co-chair), Chandra Krintz
Title: "The P's of Data Management System"
Global-Scale Data Management (GSDM) empowers social networks and the web to serve end-users
at billion-user scales. However, data management systems face many challenges to provide end-users with timely personalized experiences at scale.
In this talk, I explain how my research addresses the 3 P's of data management systems: Processing, Performance, and Privacy. In the first part of the talk, I present our work GPlacer, a placement optimization framework. It aims to find the optimal replica placement for geo-replicated strong consistent databases among the available edge and core datacenters. In the second part of the talk, I present Cache on Track (CoT), a stream processing framework that identifies heavy hitters and caches them at the edge of a distributed caching infrastructure. CoT intends to cache local trends at every front-end server to accelerate serving end-user requests. Finally, I briefly present our works LocBorg and Pharos that address different data privacy challenges at different scales.