Extending the Capabilities of Mobile Devices for Online Social Applications through Cloud Offloading

Abstract

Handheld devices are becoming an attractive option for users to interact with their social network, through online social applications. We are witnessing a rapid adoption of smarter devices all around us, which brings with it orders of magnitude in heterogeneity. Thus, researchers in the field of distributed systems are faced with new challenges: How to optimize performance for devices that are so diverse in terms of energy consumption, processing power and communication capabilities? My PhD research focuses on this challenge, adopting techniques for offloading operations from mobile to more powerful cloud-based infrastructure, and brings a three-fold contribution. First, we have characterized and modeled workloads of online social applications, and empirically validated them using traces of hundreds of real applications. Second, we are currently investigating offloading mechanisms, including: communication offloading, lossy performance offloading, and loss less performance offloading. We have been testing and evaluating these mechanisms with several mobile applications, measuring performance and energy consumption. Third, we will create an integrated cloud based offloading system that aims to improve the performance of online social applications. We will empirically evaluate this system using both simulations and open-source real-world applications.