Performance Tuning of a Storage System using Design of Experiments

Abstract: Performance testing is an essential but very costly activity. We will explore how to make performance testing cost effective and reusable using Design of Experiments. This will be illustrated by studying the performance testing in storage space. Many factors plays role in determining the performance of storage area network (SAN) and there are many research works has been carried to model the SAN performance to explain the effects of individual storage factors upon the throughput.

While most of these models analyze the individual factor effects (Main effects), the effects of the factor interaction (interaction effects) are not built into the model. This work discusses the suitability of Design of Experiments (DOE) methodology to study the main and interaction effects of factors influencing the SAN performance.

Profile: AB Kishor is Group Head for Platforms, Storage and Software Product testing at Wipro Technologies.
Currently he is working on developing solutions and frameworks in the system testing area with a special focus on Automation, Virtualization, Optimization and Performance Test Engineering. He has worked with multiple Platform, Storage and Software Product companies across US and Europe.