A load balancing method and device

Abstract

The invention discloses a load balance method and apparatus. The method comprises the steps of obtaining the utilization rate of each resource of a first storage device and the utilization rate of each resource of a second storage device in a current period, wherein mirror data required to be subjected to mutual backup exist between the first storage device and the second storage device; determining a first target resource of the first storage device and a second target resource of the second storage device; according to the utilization rate of the first target resource and the utilization rate of the second target resource, determining a target storage device required to be subjected to token release speed control in the first storage device and the second storage device; and for the determined target storage device, adjusting the token release speed of the target storage device in a next period according to the utilization rate of the target resource in a historical period of the target storage device and the utilization rate of the target resource of the target storage device in the current period.

Description

technical field [0001] The present application relates to the field of computer technology, in particular to a load balancing method and device. Background technique [0002] The unified storage system is a storage system supporting two storage architectures of network attached storage (Network Attached Storage, NAS) and storage area network (Storage Area Network, SAN), and the unified storage system includes multiple storage devices. [0003] In the prior art, each storage device in the unified storage system uses a token bucket algorithm for flow control. Each storage device needs to consume tokens in the token bucket when processing service requests, which can prevent the storage device from overloading due to burst data. The tokens in the token bucket are generated at a certain speed. When the storage device is in an overload state, the speed of token generation can be reduced; when the storage device is in a light load state, the speed of token generation can be increa...

AI Technical Summary

Problems solved by technology

At this time, for the storage device in the overloaded state, the speed of token generation will be reduced; for the storage device in the light load state, the speed of token generation will be increased; however, when increasing the speed of the storage device in the light load state At the same time as the token generation speed, it will lead to an increase in the business volume of backup image data, which will further increase the load on the storage device in the overloaded state, thus making the token regulation of the storage device in the overloaded state invalid

[0005] In summary, since the token generation speed of each storage device is individually regulated, the load among the various storage devices in the entire unified storage system is unbalanced

Images