Load Balancing Mechanisms in Distributed Computing

It is observed that very few algorithms have been developed for heterogeneous environments. Most of the Dynamic Load Balancing (DLB) algorithms available for heterogeneous environments have considered processor speed of the participating nodes. Few algorithms have taken into account the other heterogeneity factors also but due to their complexity, the algorithms are infeasible. The DLB algorithms proposed in this book incorporate most of the untouched heterogeneity factors which are highly sensitive to individual nodes. The book introduces a generic DLB model for scheduling and load balancing in clusters, routers and multi-core processors. The book also incorporated Linear Programming (LP) in load balancers. Simulation results show that application of LP has improved the performance of load balance in an unstable system. The work in the book has also emphasized the importance of open source software and proposed open source software at frontend (load balancer) as well as at backend. The book proposes an adaptive load balancing model for even workload distribution among the cores of symmetric multi-core servers too.