Article Preview
Top1. Introduction
Cloud computing service providers are providing commercial and reachable services to customers’ as per requirements. It has been analyzed that organizations are migrating to cloud for variety of services which increases the difficulty in resource management that is challenging to achieve. Customer expectations are to get services on reasonable cost and in time. Thus, service providers must ensure to provide services based on customer needs, to meet quality of service (QoS). Proper allocation will lead to higher resource availability and increase the capability to meet customers' expectations. The appropriate workload distribution on resources is difficult due to user unpredictable demand. Therefore, resource distribution and workload management are research challenges in cloud computing that impact on utilization and performance. Research surveys indicate that cloud resource allocation and management is becoming complex day by day due to tremendous demand of cloud services (Mustafa et al., 2015). Therefore, to meet higher performance in cloud computing, an efficient resource allocation and scheduling technique should be developed. Optimizing resources in system will provide solutions to considering the resource with respect to demand and other aspects. In this regard, an efficient policy to conduct the resource according to availability and scalability to ensure the expected performance (Zhan et al., 2015) is presented. Resources have been allocated to tasks, by identifying the execution cost and time through heuristic strategy. Further, resource scheduling has been done based on availability of Virtual Machine (VMs) to improve reliability, which is the focus (Pietri & Sakellariou, 2016).
The proposed, efficient resource allocation and scheduling (ERAS) technique manage resources using swarm optimization. The technique was inspired by nature that contains intelligent agents, suitable for complex and distributed environment. These agents are generated dynamic and collective solution through intercommunication process. The designed ERAS policy is based on ant colony optimization (ACO), which comes under swarm intelligence. It is an extension of ACO that performance allocation and scheduling based on resource availability. Optimal solution has been made for allocation that has a convergence process. It is similar to foraging behavior of ants that laid pheromone on the ground. The pheromone values help to generate optimal route through intercommunication. The implementation of the policy is done on cloud simulated environment and validity of results is evaluated based on QoS parameters and VMs. To minimize cost and time ERAS is planned in some phases as shown in Figure 1, to pinpoint the exact provision: 1) Task workload identification. 2) Resource availability. 3) ACO based initial feasible solution. 4). QoS based optimal solution through feasible solution intercommunication 5). Scheduling apply based on allocation decision; resource updating. Further, results are compared with traditional policies in terms of QoS.
Figure 1. Phases of efficient resource allocation and scheduling (ERAS) policy
The research work is accomplished with considering challenges for best VMs allocation as per user requirements, by reducing execution cost and time, which provides services based on customer expectation. The designed policy guarantees the cloud resource management and services by efficient allocation and scheduling through optimal solution. The paper structure is as follows: section 2 depicts cloud computing system, section 3 surveys on resource allocation and scheduling along with ACO, contribution of designed work, section 4 presented ACO based resource allocation and scheduling framework, section 5 is about simulation based experimental work and results analysis, and finally in section 6 conclusion and future work are drawn.