Power-Efficient Provisioning for Online Virtual Network Requests in Cloud-Based Data Centers

Abstract

A cloud computing paradigm enables users to access services, applications, and infrastructure resources by using thin clients anywhere and at any time. In this paradigm, multiple users can share cloud infrastructure resources. The application or service requests from a user can be abstracted as a virtual network (VN) request and can be submitted to the cloud-based data centers. How to map a VN onto the cloud infrastructure network is a challenging issue in cloud resource provisioning. Thus, efficient mapping techniques that intelligently use the resources of cloud infrastructure are important and necessary. Current research on VN mapping and design focuses on resource-efficient VN mapping or cost-efficient VN mapping. However, there is another important issue in cloud-based data centers that we must pay attention to, i.e., the amount of power or energy that is consumed by a data center. The power consumption in data centers can be a significant percentage of the total power consumption, and it not only leads to a higher data center operating cost but also contributes to carbon emissions and the greenhouse effect. In this paper, we propose a power-efficient resource provisioning technique in cloud-based data centers while complying with service level agreements. We first model a power-efficient VN provisioning problem as a mathematical optimization problem, with the objective of minimizing the power consumption by employing mixed-integer programming. We then propose a heuristic algorithm to efficiently solve this model since this optimization problem is NP-hard. We validate and evaluate our framework and algorithm by conducting extensive simulations on different cloud infrastructure networks under various scenarios. The simulation results show that our approach performs well.Index Terms-Cloud computing, data centers, embedding, power efficiency, provisioning, virtual network (VN) requests.

Energy cost minimization with job security guarantee in Internet data center

“…Sun et al [38] proposed a power-efficient resource provisioning technique while meeting the service level agreements in cloud data center.…”

Providing a Cloud Broker-based Approach to Improve the Energy Consumption and Achieve a Green Cloud Computing

“…GreenCloud consist of several components such as Monitoring Service, Migration Manager and Managed Environment. Another power-efficient resource provisioning technique in cloudbased data centers proposed by [18] in 2013. Since optimization of power-efficient virtual network provisioning is NP-hard, they propose a heuristic algorithm in cloud-based data centers.…”

Virtual network embedding for power savings of servers and switches in elastic data center networks

“…The connection probability is 50% 2) for a pair of VMs, and the spectrum requirement of a VLP has 4-8 spectrum slots. The benchmarks are load-balancing-oriented VN embedding algorithm (LB in short) [2] and VN embedding via sleeping servers (SS in short) [8]. We consider 4 performance metrics: total power consumption (including servers and switches), total switch power consumption, average power consumption per VN embedding request, and average blocking rate (ABR).…”