Mobile Application Offloading

New applications are putting increased strain on the limited processing, memory, and energy resources within mobile devices. Offloading to local servers or the cloud is a widely proposed solution for coping with these constraints. However, existing offloading systems have one or more shortcomings that have hindered their actual adoption: lack of security, tuned for a single goal, poor automated partitioning, and dedicated offload resources. We propose a framework that bridges these gaps and makes offloading a reality, at least in enterprise settings.

We argue for an offloading framework that takes advantage of diverse resources in enterprise networks (idle desktops, dedicated servers, and remote clouds) to execute applications offloaded from wireless mobile devices. All decisions of what and where to offload are made from a central controller, ensuring decisions adhere to administrator specified objectives and security policies. Making decisions with a network-wide view--which includes multiple types of computing resources, the security context of applications and data, and knowledge of each device's performance and energy goals--allows our framework to leverage all possible computational capacity and meet the individual needs of each device and application. Our goal is to make mobile application offloading work for you.