Server cluster service traffic control method and related equipment

Abstract

The invention discloses a server cluster service traffic control method and related equipment. The method includes the following steps: obtaining historical service traffic of each service operated ona server cluster; respectively inputting the historical service traffic of each service to the own matched traffic prediction models, and obtaining predicted service traffic that is output by each traffic prediction model and indicates each service at the to-be-predicted time, wherein each service is in one-to-one correspondence with each traffic prediction model; calculating the sum of the predicted service traffic of various services to obtain the predicted cluster network card traffic at the to-be-predicted time; and if the predicted cluster network card traffic exceeds the rated upper limit traffic of the server cluster, performing traffic control on the services of the server cluster at the to-be-predicted time according to a preset manner. According to the technical scheme providedby the invention, on the basis of ensuring the normal operation of various services, the processing capability of the server cluster can be fully played, and thus the waste of resources can be avoided.

Description

technical field [0001] The invention relates to the field of computer technology, in particular to a method for controlling service flow of a server cluster and related equipment. Background technique [0002] Server clustering refers to the concentration of multiple servers to perform the same service, which appears to the client as if there is only one server. The cluster can use multiple servers for parallel computing to obtain high computing speed, and can also use multiple servers for backup, so that any server failure does not affect the normal operation of the entire system. [0003] A server cluster generally runs multiple services, and its current cluster network card traffic is the sum of the current business traffic of each service. Since the number of servers in the server cluster is generally fixed, that is, the maximum traffic of the cluster network card that the server cluster can support (hereinafter referred to as the rated upper limit traffic) is fixed, bu...

AI Technical Summary

Problems solved by technology

Since the number of servers in the server cluster is generally fixed, that is, the maximum traffic of the cluster network card that the server cluster can support (hereinafter referred to as the rated upper limit traffic) is fixed, but in the actual operation process, the current cluster network card traffic of the server cluster may exceed the rated In the case of the upper limit traffic, although the carrying capacity of the server cluster can be maximized at this time, it is very likely that the business traffic of hot business will occupy a large amount of resources of the server cluster, causing other businesses to be severely delayed or even unable to respond.

[0004] Therefore, in the prior art, the flow of each business is generally limited, so as to avoid the risk that the server cluster will be overwhelmed by the hot business. Limiting the peak traffic of the business, the processing capacity of the server cluster cannot be fully utilized, resulting in a waste of resources

Images