Resource-Aware Dynamic Load Balancing System in Heterogeneous Distributed Computing Systems through Swarm Intelligence

Abstract

Distributed computing is a very common term in the field of computer science and a very general computing paradigm which is being used by almost all the scientific and industrial computing applications. Distributed services are generally deployed on a server somewhere at some other geographical region and its consumers are spread across the world. Now a days almost all the commercial applications consumes distributed services in some forms to have a set of different consumer services to their customers on their platform. As soon as this kind of computer applications started to come into the commercial market, the number of its users started to increase exponentially over the time. By the time it became very difficult to handle and respond to its numerous users at the same time which leads to a high risk of losing the customers, if they are unable to provide services with a minimal possible response time. If an application has comparatively less features but its response is fast, then it became a widely used product in the market. Hence almost all the organizations invest a separate large amount of money annually on the maintenance of its computer applications and specially to boost its performance and response. Even though there are still many challenges in this field, since organization has a limited computing and networking resources to provide their services to its users which are in millions in number. Hence, it has been always a major area of research for many engineers and scientists from all over the world to implement and deliver an algorithm that would be able to solve all these types of problem and provide the services up to the customer satisfaction. In the field of computer science, Artificial Bee Colony (ABC) introduced by D. Karaboga [12] is one of the best optimization algorithms on the intelligent foraging behavior of honeybee swarm. Therefore, we have represented this paper with a systematic study on designing a load balancing system which is artificial intelligence enough to control the network traffic on the host server to allow only permitted number of jobs at the same time which falls under resource constraint of the server. It will work as a middleware between the front-end application and channel of heterogeneous distributed computing servers.