An effective approach of latency-aware fog smart gateways deployment for IoT services

Fog computing has become a new cutting-edge technology to bring cloud applications closer to physical IoT devices at the edge of the network. In Fog computing use case scenarios latency is the main challenge that must be addressed since mission-critical environments are mostly delay sensitive. To achieve this goal, the fog nodes placement is needed to minimize the delays in the service access layer. There is neither a common fog computing architecture nor a generally accepted concept of fog nodes, how and where to deploy fog nodes or how it supports the real-time Internet of Things (IoT) service execution. Edge devices such as the switch, router, gateway, mobile phones, smart car etc., are the candidates for deployment of fog nodes but the deployment differs according to the application. In this work, we have taken gateways as candidates for fog node deployment. The gateway collects data from smart sensors, but it does not have any pre-processing or decision-making capabilities. Therefore, the gateway is made smarter with Fog capabilities and named as Fog Smart Gateway (FSG). The processing of IoT traffic is taken care of by Virtual Machines (VMs) facilitated by distributed Fog nodes. We optimized the number of fog nodes for deployment to reduce the total latency induced by traffic aggregation and processing. Our results show that the optimal deployment of fog nodes in the IoT network could yield a reduction in latency compared to processing IoT data in a conventional cloud system. (C) 2019 Elsevier B.V. All rights reserved.