Physical host load detecting method based on CPU (Central Processing Unit) heartbeat amplitude

Abstract

The invention relates to a physical host load detecting method based on CPU (Central Processing Unit) heartbeat amplitude. The method comprises the following steps: setting accordant monitoring parameters for all physical hosts, and setting activation judging moments of all physical hosts; and all physical hosts receiving detection beginning instructions, judging activation, reading CPU usage amount history record sheets (H), and calculating estimated load values (MH) and actual load values (MR), wherein when MR is larger than MH, the physical hosts are in an overload state, and otherwise, the physical hosts are in a normal state. The method can quickly and accurately detect the load state of the physical hosts, and then reduces main energy consumption indexes of the physical hosts and enhances performances of the physical hosts, and finally reduces operation, maintenance and management costs of the whole system.

Description

technical field [0001] The invention relates to a load detection method of a physical host based on the CPU heartbeat amplitude, which is an improvement aimed at the dynamic optimization mechanism of the resources of the physical host, and belongs to the technical field of computer networks. Background technique [0002] As far as the applicant knows, in the current cloud computing environment, the core idea of ​​physical host load detection methods is based on the statistical value of center deviation. For example, the currently most influential interquartile range (IQR) method belongs to the physical host load detection method based on the statistical value of center deviation. Its general process is: rearrange a series of original values ​​in ascending order, and then The sequence obtained by sorting is divided into four equal parts, and finally the difference between the value on the third quartile and the value on the first quartile is taken as the basis for judging the...

AI Technical Summary

Problems solved by technology

[0003] However, the main problem of this method is that it only considers the average load distribution of the physical host, but fails to take into account the jitter factor during the operation of the physical host, which makes it difficult for the existing methods to quickly and accurately detect the unstable operating load. physical hosts, and make the main energy consumption indicators of the physical hosts higher when the system is running (such as the actual number of virtual machines migrated, the proportion of virtual machine performance degradation caused by the migration operation, the selection time of the target virtual machine to be migrated, and the average actual migration time of the target virtual machine value), resulting in increased energy consumption of the physical host and decreased performance, which reduces user experience and ultimately increases the cost of operation, maintenance, and management of the entire system

This method uses the directly monitored parameters of the physical host as the judgment basis, and can solve the technical problems of the aforementioned existing physical host load judgment methods to a certain extent, but the method still has the following main problems: (1) Determining the IO throughput factor The IO throughput of all physical hosts needs to be taken into account, and the IO throughput factor itself is a relative ratio, which determines that this method is only applicable to systems using homogeneous physical hosts with the same load capacity. For physical hosts with different load capacities, it is difficult to judge whether the physical host is overloaded; (2) The load index needs to be determined by the three variables of CPU utilization, memory utilization, and IO throughput factor, and the load index needs to be determined based on the latest N consecutive load index Compared with the preset thresholds, although this can prevent misjudgment to a certain extent, it will inevitably increase the complexity of the judgment process, which will undoubtedly reduce the speed of judgment, and it is difficult to satisfy the cloud computing system with a huge amount of calculation. The requirement to quickly find out the overloaded physical host is also not conducive to reducing energy consumption; (3) For this method, the preset threshold is one of the key points to judge whether the physical host is overloaded, but the patent application document does not explain How to set the threshold, and obviously the threshold cannot be set arbitrarily. It takes repeated practice and research to get a suitable threshold and make the judgment result consistent with the facts. That is to say, the patent application document cannot be related to the threshold setting. teaching role

Images