If you are tackling many servers in the data center, saving a small percentage of those servers can bring significant return. In this training session, participants will learn to identify performance opportunities at scale and how this can be different from optimization on a single system.
Large-scale software deployments in the data center present several big challenges, including how to measure software performance in the data center and how to optimize software for resource management.
This training will show a few techniques developed at Alibaba to tackle some of these challenges. Kingsum Chow and Yingying Wen will cover several important considerations to avoid making mistakes in performance data analysis.
Key takeaways from this session include:
- Avoiding common mistakes in interpreting system performance data
- Interpreting CPU utilization and capacity correctly
- Drawing the correct conclusion by discovering Simpson’s paradox in the data sets
- Considerations in making decisions that scale
- Kingsum Chow, Kingsum Chow, Scientist, Software Hardware Co-Optimization, Alibaba, Zhejiang, China
- Yingying Wen, Research Intern, Alibaba Zhejiang, China
Topics: Performance Engineering
Time and Day: 1:45PM Monday, February 10