Dynamic deployment and reconfiguration of intelligent applications in mobile cloud computing with context-driven probabilistic models

Today’s mobile devices with advanced computing and storage resources are encouraging a whole new class of applications geared towards context-aware intelligence. However, for the continuous processing of context-processing algorithms, interactive human-computing interfaces and augmented-reality experiences, these devices lack the resource demands. Most of these intelligent applications rely on the cloud paradigm for on demand computing, memory and storage resources. While combining both computing paradigms, finding the best strategy to deploy and configure intelligent applications, is not straightforward. In this paper, we analyse the challenges and requirements for the dynamic deployment of intelligent applications in such a federated setting. Additionally, a framework that leverages Dynamic Decision Network (DDN) is presented for decision making. The DDN-based models deal with the presence of uncertainty and the partial observability of the context information, as well as the temporal effects of the decisions to ascertain the Qualityof- Service and Quality-of-Context requirements. Our initial experiments with the framework demonstrate the feasibility of our approach and potential benefits to automatically make the best decision in the presence of a changing environment addressing the runtime variability.status: publishe