Adaptive Mobile Cloud Workflow Management System for Service Discovery in Proximity-Based Internet of Things

Abstract

In the near future, Industrial Internet of Things can help people in realizing the smart cities and urban sensing scenarios by providing various useful spatial information. Based on real-time resource discovery and data retrieval technologies, mobile devices can continuously interact with surrounding things and provide real-time content mashup services. However, continuous resource discovery and content mashup processes can be resource intensive for mobile devices. In order to reduce the resource usage, certain tasks of the mobile application can be delegated to the utility Cloud. Nevertheless, task offloading isn't always cost-efficient and the decision to offload a process needs to be context-aware. This paper proposes a service-oriented workflow-based mobile Cloud middleware framework for balancing the task allocation between the mobile terminal and Cloud. The proposed cost-performance index scheme assists the decision-making of the workflow configuration by utilizing the fuzzy set and weight of context schemes. The evaluation consists of a data set generated by experts' preferences considering fuzzy logic and inference, on which linear regression was performed, for extracting the weights of the context. The prototype has been implemented in real devices and the evaluation has shown that the workflow system can automatically configure the task allocation based on resource availability.