Autotuning multi-tiered applications for performance

With the current popularity of web-based applications, getting the maximum possible performance from applications deployed in multi-tiered environments becomes an important task. But the large numbers of configurable parameters in today's server applications make manually tuning them for better performance difficult. In this work, we explore the autotuning approach, which is generally used to tune the performance of programs in traditional HPC settings, to tune multi-tiered web applications. Our approach is based on OpenTuner, a framework that can be used to build autotuners to search through a configuration space for an optimal configuration. However, the wide variations and the dynamic nature in the runtime environment, such as network congestion, variations in demand, possible node failures and changes in workloads pose a significant challenge for this approach. In this work, we explore offline tuning techniques to overcome the challenges of autotuning multi-tiered web applications. We present the results of offline autotuning experiments that tuned benchmark applications for multiple performance goals. We show that 20%-25% improvements in response time and throughput can be achieved through our offline autotuning approach. We also present a way of reducing the tuning time by pruning the configuration space. In addition we identify the parameters in web servers that contribute most to performance. We also discuss the shortcomings of offline autotuning methods.