| Summary: | 碩士 === 東吳大學 === 資訊管理學系 === 99 === Recently, the rising of a wave of Green IT trend of environmental protection, virtualization is a key solution to achieve emission reduction. Virtualization is designed to isolate the system resources such like computing power, memory space, and application. Also, it can make the resource usage more efficient and dynamically allocate for utilization. And, Virtual infrastructure is to implement virtualization layer on the host system, therefore the server can concurrently running multiple virtual machines (VMs) and also dynamically share the host resources between these VMs. The management of system resources response for the dynamic changing on application loads which includes scheduling and monitoring of resources such as the physical CPU, memory, and storage space. Also, it can dynamically balance host’s workload and live migration of VMs. This research is based on RedHat Virtual Machine Manager System. We propose and design automatic resource management called ARMS (Automatic Resource Management System). The ARMS employ scenario modeling director (SMD) to pre-define scenario models for automatically allocating resources via using automatic resource allocation scheme (ARAS) to enhance system resources utilization and improve the operation efficiency of VM.
The primary features of ARMS will cover: (1) to manage the power on, power off, suspend and resume of a VM and perform create, remove, edit and clone a VM; (2) to monitor and analyze the resource usage on physical servers and the workload of each VM; (3) to perform live migration methods of VM for automatic balance workload. The SMD can be employed to enable automatic resource allocation settings and make it easier to configure and manage. System administrator can directly take advantage of the SMD to apply to a VM, and trigger the ARAS to execute automatic resource allocation to enhance the host resource utilization, and ensure system performance improvement to undertake for the system operational needs. Therefore, we can significantly reduce the management overhead and enhance the operational efficiency of the VM.
|
|---|
