An Augmented Load-Balancing Algorithm for Task Scheduling in Cloud-Based Systems

Task scheduling in the cloud offers many advantages to cloud providers and users, such as managing cloud computing performances and maximizing resource utilization. However, the load might not be balanced among the multiple data centers leading to some servers being overloaded while others are idle or barely working. This paper proposes an augmented load-balancing algorithm (ALA) inspired by particle location-based search system and the Artificial Bee Colony's (ABC) memory mechanism. The search system is modified by adding the best response time criterion, best path and a data center level-based distribution system to ensure an even load handling. In contrast with the ABC and Particle Swarm Optimization (PSO) algorithms, the (ALA) takes into account the number of virtual machines (VMs) per host and the response time of each data center when scheduling the given tasks. The proposed algorithm is evaluated against other well-known techniques with a different number of experiment using the designed system model proposed. The experiments results show that (ALA) distributed the load as equally as possible and kept the system balanced having an improved response time and time.