Improved PC Based Resource Scheduling Algorithm for Virtual Machines in Cloud Computing

The existing resource scheduling algorithms for virtual machines usually use serial job deployment ways which easily lead to the job completion time overlong and the system load unbalance. To solve the problems, an Improved Potential Capacity (IPC) based

  • PDF / 1,036,028 Bytes
  • 11 Pages / 439.37 x 666.142 pts Page_size
  • 69 Downloads / 241 Views

DOWNLOAD

REPORT


Abstract. The existing resource scheduling algorithms for virtual machines usually use serial job deployment ways which easily lead to the job completion time overlong and the system load unbalance. To solve the problems, an Improved Potential Capacity (IPC) based resource scheduling algorithm for virtual machines is proposed, which comprehensively considers the overall job completion time and system load balancing, and applies a new metric to dynamically estimate the resource remaining capacities of virtual machines, and thus reduce the inexact matching between jobs and virtual machines. A batch job deployment method is also proposed to execute the batch job deployment. Many simulation experimental results show that the proposed algorithm can effectively decrease the overall job completion time and improve the load balancing of a cloud system. Keywords: Cloud computing

 Virtual machine  Resource scheduling

1 Introduction In modern data center, virtualization technology [1] has played an important role in simplifying resource management, integrating server capability and improving resource utilization. It has become the key supporting technology in cloud computing systems. Virtual machine resource scheduling [2] is one of the core techniques in this field, how to design the resource scheduling algorithms for virtual machines and thus ensure the load balancing of system and improve the user experience has been a hot research topic. Most resource scheduling algorithms for virtual machines apply serial ways to deploy user jobs, which can cause the higher resource updating frequency, the longer of the overall job completion time and the poor user experience. Furthermore, serial job deployment usually uses greedy strategy and cannot achieve the global optimum. This may lead to the serious load imbalance and affect the overall system performance. To solve these problems, researchers have proposed many resource scheduling algorithms for virtual machines. © Springer International Publishing Switzerland 2016 Y. Wang et al. (Eds.): BigCom 2016, LNCS 9784, pp. 321–331, 2016. DOI: 10.1007/978-3-319-42553-5_27

322

B. Qiao et al.

Liu et al. [3] presented two new metrics, Balance Capacity (BC) and Potential Capacity (PC) for virtual machines and used greedy strategy to deploy batch jobs serially, although it makes virtual machines have better scalability, it also can lead to the overall job completion time longer and load imbalance of a cloud system. Zhang [4] proposed a resource scheduling approach for virtual machines based on OpenStack, which can reduce the unbalance of resource and the consumption of power, but the algorithm don’t consider the dependencies of the virtual machines. Minarolli and Freisleben [5] presented a resource management algorithm based on artificial neural network which reduced the number of active servers, but this method mainly focuses on energy consumption, and do not consider load balance and user experience. [6] proposed a random resource scheduling algorithm which has good load balancing. But