ANN Based Estimation of Reputation of Newcomer Web Services in Fog Computing

Fog computing is a distributed computing model which stores and computes at the edge of the network with the reason of reducing the network overload and process the gathered information with low latency. By doing so there is improvement in the response time and the system performance. Fog environment consists of huge number of IoT devices, fog nodes and cloud service providers. The fog nodes are providing services to IoT devices and the selection of required web services in such a heterogeneous environment is a complex issue. The Web Service Recommendation System (WSRS) help IoT devices to choose appropriate service provided by fog nodes based on reputation of the its services. In service based computing environments, reputation of web services is used for recommending services. The performance of recommendation system depends highly on evaluation of the initial reputation values failing which exposes the system to various threats like whitewashing and Sybil attacks. Although the reputation management plays a major role in recommending services of fog nodes to IoT devices, the evaluation of the reputation of newcomer services without feedback ratings is challenging task. Hence this paper presents Artificial Neural Network (ANN) based estimation of reputation of newcomer web services in fog computing. The results obtained from the experiments carried out on a collection of web services in the real world suggest that the approach provided is effective and accurate in estimating the reputation of newcomer services.