System Performance Analysis at Scale

Speaker: Prof. Kingsum Chow
         School of Software Technology
         Zhejiang University

Title:  "System Performance Analysis at Scale"

Date:   Monday, 15 May 2023

Time:   4:00pm - 5:00pm

Venue:  Room 4504 (via lift 25/26), HKUST

Abstract:

When tackling many servers in the data center, saving a small percentage
of servers would bring significant return. We will describe how we
evaluate performance at scale, and how it is different from optimization
on a single system.

The emergence of large-scale software deployments in the data center has
led to several challenges: (1) measuring software performance in the data
center, and (2) evaluating performance impact of software or hardware
changes. We will highlight a couple of problems that may lead to wrong
conclusions. We will present a sketch of our solutions


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

In March 2023, Kingsum joined the faculty of the School of Software
Technology, Zhejiang University after about 30 years of working in the
industry.  He received a PhD degree from the School of Computer Science
and Engineering, University of Washington in 1996. Since then, he has
worked for Intel in USA and Alibaba in China. He received the titles of
principal engineer, chief scientist, and senior principal engineer from
the two companies he worked for. He published 127 papers and 28 patents.
During the years he worked for Alibaba and Intel, he delivered results by
collaborating with technologists from top hi-tech companies such as
Amazon, AMD, Arm, Ampere, BEA (acquired by Oracle), Google, IBM,
Microsoft, Oracle, Siebel (acquired by Oracle), Sun (acquired by Oracle)
and Tencent. He led project Apollo in the collaboration between Intel and
Oracle in the 2015 launch of Oracle Cloud, announced by Oracle and Intel
CEOs in the Oracle OpenWorld Keynote. He represented Alibaba in the
election for JCP EC (Java Community Process Executive Committee), the
highest-ranking Java authority in the world. Alibaba is still the only
company in China that has achieved this status. While at Alibaba, he led
the development a performance analysis platform called System Performance
Estimation, Evaluation and Decision (SPEED) to tackle the analysis of
system performance analysis at scale.

Outside of work, Kingsum has fun playing with LEGOs in after school
programs. He coached multiple FIRST LEGO League and FIRST Tech Challenge
teams from 2005 through 2016, through the Intel Volunteering program. He
had a great deal of fun traveling with students to different cities for
robotic competitions and occasionally took some awards in world
championships. In fact, Kingsum learned practical machine learning from
the middle school and high school students while they were getting their
robots to work. Then, he applied what he learned from the kids to work.
During the pandemic, Kingsum enjoyed hiking in China while listening to
audio books.

Kingsum can be reached on LinkedIn: https://www.linkedin.com/in/kingsumchow