Optimation of Xen scheduler for multitasking

The rise of cloud computing brings virtualization technology continues to heat up. Based on Xen´s I/O virtualization subsystem, under the virtual machine environment which has multi-type tasks, the existing schedulers can´t achieve response with I/O-bound tasks in time. This paper presents ECredit scheduler combined complexity evaluation of I/O task in Xen´s I/O virtualization subsystem. It prioritizes the I/O-bound task realizing fair scheduling. The experiments show that the optimized scheduling algorithm can reduce the response time of I/O-bound task and improve the performance of the virtual system.