Mobile Cloud Computing is the intersection of mobile computing devices (cellphones, tablets, etc.) and familiar cloud computing concepts; mobile devices can see significant benefits from leveraging the cloud, from increased computational speed and extended battery-life (generally a major concern for mobile devices), to the ability for a resource poor device to run resource intensive mobile applications. The presentation therefore will be about cloud computing as it applies specifically to mobile devices. The vast majority of the discussion is applicable to general cloud computing (and thus of broader interest to the class), but there will be a focus on the challenges pertinent to a mobile device using cloud computing (some cursory background on necessary included topics, for example, mobile network tower handoff, will be provided as needed). Such challenges include, but are not limited to: latency and connectivity (both of which effect seamlessness and QoS); and computational offloading which has many subproblems such as how much of the/which of the computations do we offload, why, and what are the structural/architectural limitations thereof. Further discussion will explore proposed/tested solutions to these specific issues, while also elaborating on some of the measurable benefits of adopting cloud computing for mobile devices: device-battery savings, faster processing, and the ability for resource poor devices to run resource intensive applications. Many, quite possibly all, of the issues discussed won’t have “given”/standard answers as they are areas of open research (especially where the limitations imposed by mobile devices are applied), but the discussion will be backed with hard data and specific approaches to these issues drawn from the available research literature. Expected topics of discussion relating to the problems presented by mobile cloud computing will include cloudlets (vs cloud frameworks) with regard to seamlessness and QoS, and methods of offloading computations (theory/algorithms for execution offloading: why dynamic not static code partitioning, minimizing state transfer costs, and other possible considerations).
The complete presentation is available here: MobileCloudComputing