Load balancing algorithms in heterogeneous web cluster

The load balancing is dedicated for distributed and parallel systems to manage the web cluster. It distributes the load among web servers using several scheduling algorithms. This paper investigates a detailed comparison of some HAProxy algorithms in heterogeneous web cluster as well as the meta-heuristic methods that inspired by the insects' colonies behavior such as Ant Colony Optimization algorithm and Honey Bee algorithm. The basic idea of meta-heuristic methods is to solve combinatorial optimization problems.

I. Introduction

In present distributed systems, the essential element is the load balancing. Its proper scheduling of tasks is a crucial optimization concern. Load balancing is the process of assigning the total loads to all nodes of the computing system. The major focus of reassigning load is to afford minimal requests' imbalance among the web cluster and maximal resources utilization. The imbalance of loads between servers of the same web cluster ought to be inferred from the heterogeneity of cluster. Thus, in heterogeneous web cluster, the characteristics of servers are different from the specification, capacity and type.