Distributed Algorithms for Resource Allocation and Load Balancing

Abstract

Modern computing is increasingly using distributed systems, and improving their load distribution and resource allocation is essential for obtaining optimal performance. Distributed algorithms for resource allocation and load balancing employ a variety of techniques, including heuristic-based, optimization-based, and machine learning-based ones. In this work, we present a review of distributed load-balancing and resource-allocation approaches. We explore the difficulties in developing efficient algorithms and emphasise the need of meticulously analysing and contrasting various algorithms in light of the requirements of a certain system and workload. Additionally, based on the concept of particle swarm optimisation, we present a distributed method for load balancing and resource allocation in cloud computing environments. Our suggested method tries to reduce the average task waiting time while simultaneously maintaining some semblance of resource parity among nodes. By putting our technique through its paces in a simulated cloud computing environment and examining the outcomes, we compare it against cutting-edge algorithms. Our research demonstrates that our suggested technique has the potential to greatly improve system performance by reducing the typical amount of task waiting time and ensuring that the load is distributed evenly among nodes. This shows how particle swarm optimisation may be used to create efficient distributed load-balancing and resource-allocation algorithms.