A Load Balancing Approach Using Binary Search

Allocating user requests in the cloud data centre is a complex optimization challenge referred to as NP-hard. The cloud architecture is assigned a certain load based on the cloud infrastructure and user requests, which might be under laden, overladen, or balanced. Efficiently distributing jobs, whether dependent or independent, among virtual machines (VMs) is a crucial component of task scheduling in cloud computing. Load balancing (LB) involves identifying nodes that are either overloaded or underloaded and redistributing the workload among them for optimizing resource utilization, enhancing performance, and assuring high availability. This paper discusses the issues related to the inefficient use of cloud resources and different load balancing strategies. This study also discusses the comparative study of various LB techniques based on task assignment on VMs and various search techniques based on the time complexity (worst and average case) and space complexity. This study introduces a search method for finding physically heavy machines that involves job load balancing using binary search. The method has a space complexity of O(1) and a time complexity of O(log n). Upon receiving a new request, the load balancing process can be enhanced by doing a binary search on a sorted list of virtual machines based on their descending order of load, rather than arbitrarily selecting a VM or using a basic Round-Robin method, to efficiently identify an underutilized VM. If the system is underutilized (i.e. below a specified threshold), send the request to a balanced physical host after validating the load. If the system is overladen, search the lower or upper half of the list according to the load comparison. This process continues until an underloaded virtual machine is identified or the search is completed. The suggested study aims to enhance the effectiveness and efficiency of load balancing systems. It could also benefit academics and professionals in the field of distributed computing who are working to enhance the efficiency of load balancing systems.