Omni-Stream: High Performance Stream Processing in All Ways

Speaker: Dr. Shuhao ZHANG
         Singapore University of Technology and Design

Title:  "Omni-Stream: High Performance Stream Processing in All Ways"

Date:   Monday, 6 March 2023

Time:   10:00am - 11:00am HKT

Zoom Link:
https://hkust.zoom.us/j/465698645?pwd=aVRaNWs2RHNFcXpnWGlkR05wTTk3UT09

Meeting ID: 465-698-645
Passcode: 20222023


Abstract:

Stream processing has been a thriving research field for over 20 years and
is currently in its prime. Recent years have seen the successful
deployment of many stream processing systems at scale. However, their full
potential is hindered by fundamental issues such as poor utilization of
cutting-edge hardware and inadequate support for new stream applications
In this talk, I will present our recent research in two aspects: (i)
building hardware-aware stream processing systems that efficiently use
hardware such as NUMA, GPU, and mobile CPUs, and (ii) developing
specialized stream processing systems to support novel applications like
transactional stream processing and stateful stream compression.


******************
Biography:

Shuhao Zhang is an Assistant Professor at Singapore University of
Technology and Design. Previously, he was a postdoctoral researcher at
Technische Universität Berlin under Prof. Volker Markl. He earned his PhD
and BEng from National University of Singapore and Nanyang Technological
University, respectively. His research focuses on designing database
systems and big data processing frameworks for modern hardware (NUMA
multicore, GPU, mobile CPUs) with a special interest in stream processing
systems. His work has been published in top conferences/journals including
SIGMOD, VLDB, ICDE, SC, USENIX ATC, IEEE TPDS, and IEEE TKDE. He has
open-sourced most of his stream processing and database systems, including
BriskStream (ICDE'17, SIGMOD'19), TStream/MorphStream (ICDE'20,
SIGMOD'23), and OmniDB/FineStream (VLDB'13, ATC'20, TPDS'21).