Recently, resource management is the major issues in cloud computing (CC) environment because of dynamic heterogeneity of cloud computing environment. The task scheduling and virtual machines (VMs) allocation play a vital role in resources management of CC. Most of existing works for these issues aim to achieve single objective as maximizing resource utilization, load balancing, or power management. Currently, the big challenge in CC is building task scheduling and virtual machines (VMs) allocation algorithms that consider all these objectives in the same time. This problem is called task scheduling with VMs allocation multi-objectives problem which is NP completeness problem. In this paper, a new task scheduling algorithm is proposed for achieving efficient resource management based on these objectives. This proposed algorithm uses a modified genetic algorithm (GA) to find the optimal solution for choosing the most appropriate VMs for executing received tasks and their appropriate servers that will deploy these VMs. This proposed algorithm uses a matrix structure for representing the chromosome of GAS which combines the ids of tasks, VMs, and servers. Simulation results show that the proposed algorithm achieves better performance than ETVMC, TSACS, and ACO algorithms in terms of makespan, scheduling length, throughput, resource utilization, energy consumption, and imbalance degree.

Recently, cloud computing has become the most common platform in the computing world. scheduling is one of the most important mechanism for managing cloud resources. Scheduling mechanism is a mechanism for scheduling user tasks among datacenters, host and virtual machines (VMs) and is an NP completeness problem. Most of existing mechanisms are heuristic and meta-heuristic methods, developed to address a part of scheduling problem and did not consider the dynamic creation of VMs by taking into account the required resources for a user task and the capabilities of a set of available hosts. To deal with this dynamic behavior, this paper introduces a new mechanism that uses a genetic algorithm (GA) for establishing a flexible scheduling mechanism that can adapt the dynamic number of VMs based on the required resources by user tasks and the available resources of hosts. Simulation results show that the proposed algorithm can distribute any number of user tasks on the available resources and it achieves better performance than existing algorithms in terms of response time, makespan, FlowTime, throughput, and resource utilization.

No abstract