Article Preview
TopLoad Balancing algorithms are of two types: ‘static’ and ‘dynamic’ algorithms. The ‘static’ algorithms are those in which host server does not have much variations in load.
Round Robin (Samal et al., 2013) is a ‘static’ load balancing algorithm. It takes into account the principle of ‘time quantum.’ Each user request is given a time quantum and has to complete within that period of time. First Come First Serve ‘FCFS’ (Samal et al., 2013) algorithm works on the principle of a ‘first-come-first-serve’ basis. The request which come first are allocated first to the server.
Modified throttled (Domanal et al., 2013) is dynamic load balancing algorithm which moves the user request to the available resource provider. These traditional algorithms have many limitations due to dynamic workload environment. Therefore, to overcome such limitations Swarm Intelligence SI algorithms are used.
Gerardo Beni and Jin Wang introduced the idea of swarm intelligence (SI) in 1989 (Yang et al., 2017). SI describes the collective behaviour of natural and self-organized systems. SI contains agents that communicate in natural environment. Algorithms of SI optimization (Yang et al., 2017) are ant colony optimization, artificial bee colony, Firefly algorithm, cuckoo search, etc. Metaheuristics define a procedure to follow a particular path that leads to best optimization problem.
The Ant Colony Optimization (Tawfeek et al., 2015) is based on the behavior of ants. The collective extensibility of these ants working in parallel manner helps in solving various difficult tasks. However, ‘ACO’ provides slower convergence rate and increases the network overhead with the increase in ants.
Artificial Bee Colony (Gamal et al., 2017) (Babu et al., 2016) is based on the foraging behavior of bees. These bees are divided into different categories like ‘Employer’, ‘Onlooker’ and ‘Scout’ bees. The limitation of the ABC algorithm is that it provides slow convergence rate and poor performance on smaller paths.