Research On The Resource Schedule Strategy Based On Virtual Machine Live Migration

Virtualization technology as a key technology of cloud computing has been widespread concerned, and live migration technology of virtual machine is even the hot spots of virtualization technology. Because the live migration of virtual machine has a high granularity, lots of computing center that take virtual machines to achieve their resource virtualization are using live migration technology to achieve the load balance of these computing centers. But as the increase of user number and system demand, more and more servers are added to the computing center, resulting in the calculation of the energy consumption is growing higher. The energy consumption is becoming a major problem that plagues the cloud computing provider while only a few research focus on the load balance and energy consumption at the same time. Under this background, this paper proposes a new resource schedule strategy based on live migration that can not only meet the requirements of load-balance but also the requirement of power-saving. The main work of this paper:1. Based on the traditional centralized resource schedule framework, design an equalized schedule framework that not only has all the features of centralize framework but also has a more efficient and fault tolerance, etc.2. Design a migration trigger strategy based on load forecasting that avoids the migration due to short-term load increase, effectively reducing the times of live migration.3. Design a virtual machine selection strategy that bases the CPU usage and memory usage. Minimize the additional overhead of migration when reducing the load of the host.4. Design a destination host selection algorithm based on weighted probability, effectively avoid the negative effects that critical mass brings.The resource scheduling strategy is tested and analyzed in chapter six, the experimental results show that the proposed strategy has a good performance both on load-balance and power-saving.