Michael Stumm – Continuous Deployment of Software
报告一题目: Continuous Deployment of Software
摘要Abstract: Continuous deployment is the practice of releasing software updates to production as soon as it is ready, in increments as small as possible. This practice has been adopted by many high-profile Internet firms such as Amazon, Etsy, Facebook, Flickr, Google, and Netflix that deploy software updates to production 10s, 100s, or even 1,000s times a day and programmers release software once a day on average. The practice has also been applied on mission-critical software as well as firmware.
This talk will describe the software development and deployment processes used to achieve continuous deployment, and lists critical requirements needed to make continuous deployment successful. An analysis of how continuous deployment has affected software productivity and software quality at Facebook and OANDA over a period of seven years will be presented. Finally, we describe how continuous deployment can be applied on mobile software, which is particularly challenging.
报告人简介Bio: Stumm is a teacher, a researcher, an entrepreneur and a former executive. As a computer engineering professor at the University of Toronto, he has published over 100 papers in top-tier conference proceedings and scientific journals; he holds 10 patents. His area of research is large-scale systems software and software engineering. Stumm co-founded a number of startups, including OANDA Corp., an early pioneer in online foreign currency trading, which he ran for 15 years. He is a technology consultant to a number of companies, including Facebook, and is an advisor to Singapore’s central bank and regulator, MAS. He is an active angel investor and advisor.
报告二题目：Reliability and Performance of Large-scale Distributed Systems
主讲人：多伦多大学Ding Yuan 助理教授及其团队
Software services today are built on distributed systems. This talk will cover our recent research on the improving the reliability, failure diagnosis, and performance of such systems. In particular, it will discuss three topics: (1) automated software logging to aid postmortem failure diagnosis, (2) automated failure reproduction, and (3) understanding and eliminating the warm-up overhead of Java Virtual Machine.
Ding Yuan is an assistant professor in the Electrical and Computer Engineering department at University of Toronto. His research focuses on systems software. His research result has been widely adopted by companies including Microsoft, Huawei, NetApp, and open-source software projects including Hadoop, HBase, Linux kernel, etc. Having published 8 SOSP/OSDI papers, he is the second most prolific researcher in SOSP/OSDI community in the last 6 years in the world. He received his BE and BS from Beihang University.