Hybrid load-balancing algorithm for distributed fog computing in internet of things environment

Abstract

Fog computing is a novel idea created by Cisco that provides the same capabilities as cloud computing but close to objects to improve performance, such as by minimizing latency and reaction time. Packet failure can happen on a single fog server across a large number of messages from internet of things (IoT) sensors due to several variables, including inadequate bandwidth and server queue capacity. In this paper, a fog-to-server architecture based on the IoT is proposed to solve the problem of packet loss in fog and servers using hybrid load balancing and a distributed environment. The proposed methodology is based on hybrid load balancing with least connection and weighted round robin algorithms combined together in fog nodes to take into consideration the load and time to distribute requests to the active servers. The results show the proposed system improved network evaluation parameters such as total response time of 131.48 ms, total packet loss rate of 15.670%, average total channel idle of 99.55%, total channel utilization of 77.44%, average file transfer protocol (FTP) file transfer speed (256 KB to 15 MB files) of 260.77 KB/sec, and average time (256 KB to 15 MB) of 19.27 sec.

Utilizing Probability Distribution for Selecting Optimal and Minimal Replicas to Achieving Fault Tolerance in a Distributed System

“…User interface: This interface allows users to submit their tasks to the network. The received tasks are then directed to the available network resources [22][23][24][25].…”

A survey on edge and fog nodes' placement methods, techniques, parameters, and constraints

“…Edge computing and Fog computing have the same fundamental aim and concept of bringing the storage, computing, and networking resources to the network's edge [13, 14]. However, we need to answer a second question “What are the main features of edge computing, cloudlets, and fog computing?”…”