Towards systematic mobile cloud performance analysis

The performance of mobile applications has critical impact on their adoption and success. Mobile scenarios especially in enterprise context require the connectivity to backend services that are hosted in cloud environments. Performance engineering in such a setting is challenging due to the heterogeneous influencing factors, namely limited resources on the mobile device, the available mobile network, and the characteristics of the cloud backend. In order to tackle this challenge, we propose a measurement approach based on automated experiments. By collecting performance metrics from the web frontend, traffic statistics from the device, and performance data from the cloud backend, we get an end-to-end view on the overall performance. Parameterised performance experiments allow a systematic variation in thel workload and application dimensions of interest. As a result, the influencing factors of mobile cloud performance can be clearly identified and quantified, and build the foundation for feedback to development and deployment, and subsequent performance improvements and predictions.