Service priority queuing model-based internet of things middleware for load balancing among fog computing centres

Abstract

The fog computing architectural component meets the major requirements of the internet of things (IoT) applications. Fog computing nodes are mainly used to handle delay sensitive requests generated by IoT devices. An increase in the service arrival rate at the fog computing centre leads to an overload of fog nodes. Offloading delay sensitive requests from the overloaded fog node to the cloud introduces significant communication delays. This may lead to a violation of the service deadline in the service level agreement (SLA). This paper proposes horizontal collaborative load balancing among fog nodes in close geographical proximity. The authors use the preemptive priority queuing model for the processing of delay sensitive requests. The paper introduces an efficient message exchange among fog nodes to reduce computational overhead. These messages are used for load balancing. The result of the simulation shows that the proposed approach demonstrates the superiority of the proposed mechanism in minimising the response time of delay sensitive requests. Moreover, 11% reduction in the offloading of delay sensitive requests to neighbouring fog nodes is a significant improvement over existing methods.