Resource prediction method and system and capacity management apparatus

Abstract

The present invention provides a resource prediction method and system and a capacity management apparatus. The method comprises the steps of according to a service prediction algorithm and historical data of an initial service index, predicting the initial service index, and acquiring a prediction result of the initial service index; acquiring an association factor of the initial service index, and acquiring historical data of the association factor; predicting the association factor according to the historical data of the association factor and the service prediction algorithm, so as to obtain a prediction result of the association factor; and according to the prediction result of the initial service index and the prediction result of the association factor, acquiring a prediction result of a final service index, and predicting resource consumption of a server according to the prediction result of the final service index. According to the resource prediction method and system and the capacity management apparatus, resource prediction accuracy is improved.

Description

technical field [0001] The present invention relates to the field of cloud computing, in particular to a resource prediction method, system and capacity management device. Background technique [0002] Capacity management is used to evaluate the capacity of the existing network site, that is, to collect and analyze the capacity of CPU, memory, storage, network and other resources under the current business volume scale, and predict the system based on the current business scale and business volume growth trend future support. This supports capacity scheduling or capacity expansion in advance to ensure the smooth operation of the system. Therefore, a medium- and long-term resource forecasting and estimation method is needed. [0003] At present, the resource prediction and estimation of servers or service nodes are all predicted by predicting the resources consumed by the services themselves called by the business on the server. However, the resource consumption affecting s...

AI Technical Summary

Problems solved by technology

[0003] At present, the resource prediction and estimation of servers or service nodes are all predicted by predicting the resources consumed by the services themselves called by the business on the server. However, the resource consumption affecting services is complicated and may be affected by other service calls. Therefore, the current server prediction results usually have relatively large errors and are not accurate enough

Images