Cloud-Guided QoS and Energy Management for Mobile Interactive Web Applications

In mobile interactive web applications, energy-efficient quality-of-service (QoS) scheduling involves setting a deadline for the best user experience and providing just enough performance to minimize energy. Such performance-slacking approaches require precise performance adjustment using execution time prediction. However, prior prediction approaches suffer from prohibitive training due to extensive input data and manual source code instrumentation. In this paper, we propose a cloud-guided QoS and energy management approach that eliminates the prediction overhead by offloading it to cloud resources. Our approach pre-computes per-input execution time models by profiling web applications on dedicated mobile devices in the cloud. When mobile web applications request data to servers, both the data and its execution time models are delivered to users' mobile devices. Based on the delivered models, a performance control agent on the mobile device selects an operating point to meet the response time requirement. Experimental results show that, by offloading modeling and prediction overheads, our performance-slacking approach can provide average energy savings of 22% and 39% (and up to 89%) for two different timing budgets compared to an industry-quality approach.