دسترسی بالا و مصرف انرژی کارآمد برای خدمات محاسبات ابری با زیرساخت های شبکه

The main services in cloud computing are SaaS (Software as a Service), PaaS (Platform as a Service) and IaaS (Infrastructure as a Service). Among these services, server virtualization falls into IaaS which is a service for lowering server maintenance costs. Generally, the primary aim of server virtualization has been to improve system performance by replacing multiple servers with a single server with better performance. But it carries the problem of increased cost, needed for the installation of new servers for server virtualization. In this paper, grid infrastructure is used for server virtualization in which existing servers are used rather than bringing in new servers. Server virtualization service is provided using scheduling algorithms for distributed servers or resources in grid computing. To overcome potential performance limitations that come from using existing servers, mathematical models of Meta and Sleep Servers under the grid infrastructure environment are used to provide server virtualization service with high availability.

In knowledge information society, increasingly large amounts of data are driving the advancement of systems that process them [1], [2] and [3]. The additional servers required for this advancement is rapidly increasing the maintenance costs for the information technology infrastructure. Furthermore, there are additional costs incurred, because more servers have to be installed, more space needs to be secured to house the servers, manpower has to be increased to maintain the servers, and more energy costs are needed to power the servers. In order to lower these costs needed for maintaining the IT infrastructure, many are turning to cloud computing for a solution [2], [4] and [5]. Server virtualization is an integrated technology comprised of physical and logical aspects, whose aim is to increase availability of multiple servers. Server virtualization systems such as EMC (VMware) [4], Microsoft (Virtual Server 2005), SWsoft (Virtuozzo), Sun (Solaris Container) and XenSource (Xen Enterprise3) [7] are based on integration after switching to high-performance servers from existing servers. But to take advantage of server virtualization, existing servers have to be replaced for new high-performance servers because that’s how server virtualization works: providing multiple guest OS environments using a single high-performance server. Thus a lot of costs are needed for the upgrade [4], [6], [7], [8] and [9]. In this paper, to overcome the problem of increased costs needed for the server upgrade, grid infrastructure is used. In particular, server virtualization is implemented using grid infrastructure in order to use existing servers. It also presents a scheme for increasing availability of existing servers and for reducing energy costs. The issue of lowered performance that would result from using existing servers for server virtualization is handled using Sleep Server (sub server) and Grid infrastructure. This system is also designed to reduce energy consumption. In the system, the existing servers are divided into the Meta Server group and the Sleep Server group based on the grid infrastructure (as shown Fig. 1). For Meta Server, existing servers are turned into Virtual Nodes or Virtual Servers for the grouping while for Sleep Server, existing servers are mapped to Sleep Nodes for the grouping. The Grid Infrastructure Virtualization server (GIV) is designed and implemented, which performs scheduling of jobs and resources by controlling the virtualized server system. It also performs monitoring of job list and statistical processing.

This paper presented the GIV system which has higher availability and more efficient energy consumption compared to the traditional system. It does this by using grid infrastructure to use idle resources and Sleep Server resources. That is, rather than bringing in new servers for server virtualization, the existing servers were used in conjunction with the grid infrastructure to achieve the aims. Also, an efficient energy consumption scheme is proposed in which the ACPI (Advanced Configuration & Power Interface) standard specification is used for power management of Sleep Server. In addition, Sub Servers were used to improve performance and to minimize energy consumption. In particular, a monitoring program was used so that idle resources of the server and job schedule could be viewed. In addition, statistics of system idle resources and resource usage were shown so that the benefits of the GIV system could be readily seen. Our testing showed that the GIV system had downtime of less than 5 min a year, which makes the usage about 99%, making it a highly available system.