An Enhanced Dynamic Load Balancing Mechanism for Task Deployment in IaaS Cloud
Mainak Adhikari, Tarachand Amgoth · 2018 International Conference on Computing, Power and Communication Technologies (GUCON) · 2018
Cloud computing emerges a new paradigm in a distributed environment which provides various types of computing resources over the internet based on the requirements of the user applications. Due to the increasing demands and diversity of the user applications, most of the servers are heavily overloaded which makes the whole cloud data center load imbalanced. Most of the previous works deploy the tasks to the server through some static or dynamic mechanism without concerning about the current load of the resources on the server. This may reduce the resource utilization of the servers and makes the whole network as load imbalanced. To overcome the above-mentioned challenges, we propose an enhanced dynamic load balancing mechanism, ED-LB in Infrastructure-as-a-Service (IaaS) cloud. The algorithm first selects the best-fit virtual machine (VM) instance for each task based on the resource requirements by the users and deploys the selected VM instance dynamically to the minimum loaded server on the cloud data center. The aim of the ED-LB algorithm is to utilize the resources efficiently on the servers and balance the load on the cloud data center. The proposed ED-LB algorithm is compared with various current state-of-arts-algorithms over various synthetic datasets and evaluates its performance over various metrics such as makespan, throughput, reliability, availability, evaluation of load balancing and resource utilization.