A Fine-Grained Adaptive Middleware Framework for Parallel Mobile Hybrid Cloud Applications

Mobile Cloud Computing (MCC) overcomes mobile device limitations by delegating tasks to more capable cloud spaces. Existing mobile offloading solutions generally rely on full virtual machine migration, which is coarse-grained and costly, or implementation of code offloading as part of the application logic, which greatly increases the application complexity and the associated software development costs. Some recent solutions implement fine-grained offloading, but pause the local mobile application while waiting for the offloaded code results. This leads to sequential execution and wastes local mobile resources and ignores the potential elasticity of the cloud environment. We have developed the IMCM framework to support parallel mobile application offloading to multiple cloud spaces. IMCM is finegrained, supporting application distribution at the granularity of individual components; it is adaptive, addressing the dynamicity in run-time conditions and end-user contexts; and it is fully parallel, supporting both parallel application model and simultaneous execution at mobile device and multiple private and public cloud spaces. Our evaluation results show that IMCM can improve the performance of computationally intensive mobile applications by a factor of over 50, while masking the underlying complexity of mobile-to-cloud code offloading.