Towards Energy-Efficient, Fault-Tolerant, and Load-Balanced Mobile Cloud
MetadataShow full item record
Recent advances in mobile technologies have enabled a new computing paradigm in which large amounts of data are generated and accessed from mobile devices. However, running resource-intensive applications (e.g., video/image storage and processing or map-reduce type) on a single mobile device still remains off bounds since it requires large computation and storage capabilities. Computer scientists overcome this issue by exploiting the abundant computation and storage resources from traditional cloud to enhance the capabilities of end-user mobile devices. Nevertheless, the designs that rely on remote cloud services sometimes underlook the available resources (e.g., storage, communication, and processing) on mobile devices. In particular, when the remote cloud services are unavailable (due to service provider or network issues) these smart devices become unusable. For mobile devices deployed in an infrastructureless network where nodes can move, join, or leave the network dynamically, the challenges on energy-efficiency, reliability, and load-balance are still largely unexplored. This research investigates challenges and proposes solutions for deploying mobile application in such environments. In particular, we focus on a distributed data storage and data processing framework for mobile cloud. The proposed mobile cloud computing (MCC) framework provides data storage and data processing services to MCC applications such as video storage and processing or map-reduce type. These services ensure the mobile cloud is energy-efficient, fault-tolerant, and load-balanced by intelligently allocating and managing the stored data and processing tasks accounting for the limited resources on mobile devices. When considering the load-balance, the framework also incorporates the heterogeneous characteristics of mobile cloud in which nodes may have various energy, communication, and processing capabilities. All the designs are built on the k-out-of-n computing theoretical foundation. The novel formulations produce a reliability-compliant, energy-efficient data storage solution and a deadline-compliant, energy-efficient job scheduler. From the promising outcomes of this research, a future where mobile cloud offers real-time computation capabilities in complex environments such as disaster relief or warzone is certainly not far.
Distributed data storage
Distributed data processing
Chen, Chien-An (2015). Towards Energy-Efficient, Fault-Tolerant, and Load-Balanced Mobile Cloud. Doctoral dissertation, Texas A & M University. Available electronically from