A hybrid framework using Markov decision process for mobile code offloading

With the emergence of a promising technology, Cloud Computing is spreading its usage everywhere. Mobile Cloud Computing (MCC) integrates cloud computing with mobile devices which holds a promise to run powerful and resource-intensive applications on mobile devices irrespective of its performance limitations (e.g., battery life, memory utilization and computation). Overcoming of these limitations is realized through a prominent solution, computation offloading: bypassing heavy computation to resourceful servers and receiving the outcome from these servers. A lot of research works have been done for mobile code offloading. Most of the frameworks either requires full virtual machine migration which results in coarsegrained and costly solutions, or implement fine-grained offloading requiring costly synchronization between local and offloaded codes. In both approaches, offloading of native methods are restricted only to local execution. In this paper, we have proposed a conceptual models blending the positive outcomes of both approaches using Markov Decision Process (MDP) for decision making and provides conceptual analysis with related research works.