Feedback-based service selection in ad-hoc mobile cloud computing

The collaborative efforts to overcome the challenges of Mobile Cloud Computing (MCC) has resulted into development of new portable devices, improved area cellular coverage and seamless integration of wireless data access into mobile devices. The Ad-hoc system imbibes these new advantages of mobile devices, enabling devices to act as both service provider and consumer. One main challenge in this system is the time taken to select an appropriate route to the mobile node that provides the service, especially from sets of similar nodes performing similar functionalities. Though, Quality of Services (QoS) and modified shortest path algorithm are among the solutions provided in the literature among selection approach. This approach is not totally suitable for an Ad-hoc mobile cloud system with lack of large memory space to run bogus algorithms. Moreover, QoS descriptions do not always give full details of service parameters. These challenges make service selection approach in Ad-hoc mobile cloud environment less effective. Therefore, to avoid the time consuming and expensive real-world service invocation, this paper proposes a user feedback system which uses item based collaboration filtering. This involves using user feedback dynamically to make predictions to meet user requirements. We developed a simple user feedback system for an m-Learning scenario that takes in user's service ratings after invocation. The weighted sum is scaled by the sum of similarity terms between requested service and advertised services to ensure the prediction is within the predefined range of available services. Our results show a considerable performance based on response time in comparison with our earlier approaches.