Cluster system CPU (central processing unit) distributing method and device

Abstract

The invention discloses a cluster system CPU distributing method and device. The method comprises acquiring the number of times of I / O (input / output) requests as well as the CPU utilization rate of every virtual machine; comparing the number of times of I / O requests of every virtual machine with a threshold value 1, the CPU utilization rate with a threshold value 2, and if the number of times of I / O requests is greater than the threshold value 1 and the CPU utilization rate is lower than the threshold value 2, determining a corresponding virtual machine to be a delay-sensitive virtual machine; otherwise, determining the corresponding virtual machine to be a delay-insensitive virtual machine; binding the delay-sensitive virtual machines with a first physical CPU (central processing unit), the delay-insensitive virtual machines with a second physical CPU, wherein the length of the time slice of the first physical CPU is shorter than that of the second physical CPU. Therefore, the cluster system CPU distributing method achieves effective scheduling of the physical CPUs, enhances the performance of the virtual machines and improves the resource utilization rate.

Description

technical field [0001] The invention relates to the field of network space connection, in particular to a method for distributing CPUs in a cluster system, and also relates to a device for distributing CPUs in a cluster system. Background technique [0002] With the maturity of operating system virtualization technology, the virtualization system allows multiple operating systems to run concurrently on the same physical machine, use the same hardware, and complement each other. Multiple virtual machines can form a cluster, and the cluster system has gradually become the focus of the computer industry due to its low cost, powerful computing power and robust fault tolerance mechanism. [0003] A virtual machine usually needs to use the physical CPU of the corresponding physical machine to implement I / O operations and computing operations. The physical CPU resources of the virtual machine are shared. Among them, the physical CPU takes a long time to implement computing operatio...

AI Technical Summary

Problems solved by technology

[0003] A virtual machine usually needs to use the physical CPU of the corresponding physical machine to implement I / O operations and computing operations. The physical CPU resources of the virtual machine are shared. Among them, the physical CPU takes a long time to implement computing operations, and computing I / O O operation only needs to consume a very short time. However, in order to ensure that the time slice length set by the physical CPU can be applied to the above two situations at the same time, the time slice length of the physical CPU is usually set according to the time required for the calculation operation. As a result, when the physical CPU completes the I / O operation, the time slice allocated to the operation is much longer than the time required for the operation, thus resulting in waste of time, which in turn affects the performance of the virtual machine and makes resource utilization low.

Images