Resource Selection Service Based on Neural Network in Fog Environment

Abstract

As an emergent technology in Internet of Things (IoT), the ultimate target of fog computing is to provide a widely distributed computational resources and data repository closer to the network edge providing heterogeneous systems both in terms of software and hardware. The fog system must have the capability to deal with huge number of resources and users at the same time, such increased size sometimes presents the issue of performance degradation. Therefore, the fog should be able to support adaptability, scalability, and extensibility to avoid such degradation by adopting efficiently and effectively an optimal resource selection and allocation model. As many fog users have limited interest in, or knowledge of fog middleware issues, it follows that in a large fog environment the best approach would be to have an automated system for resource selection and allocation. Such an approach eliminates the need for user intervention. This paper proposes a fog resource selection service based on neural network to perform resource selection tasks by coordinating with the metascheduler. Five different selection algorithm were used to evaluate the prediction model for resource selection. In addition to introducing a history update and management algorithm to manage and control the storage of the history log records.

LotkaVoltera and Elman Hebbian Recurrent Neural Network Cache-Based Resource Allocation in Fog Environment

“…IoT with FC involves hardware and software resource allocation; response time and busy time were enhanced; bandwidth and jitter rate was not. The neural network and meta-scheduler in [3] effectively predicted fog resources. Five algorithms were employed to identify the best help.…”