Adaptive external storage IO performance optimization method

Abstract

A method for optimizing IO performance of adaptive external storage device includes requesting to distribute priority for IO arrived at external storage device driving program and inserting it into IO request processing queue, obtaining statistic data of IO access address increment and preprocessing obtained data, calculating access address of preread data by utilizing said statistic data, calculating size of preread data starting from access address of preread data and from data used by external storage device to fetch preread data size to data buffer-storage.

Description

technical field [0001] The invention belongs to the field of performance optimization of computer systems, in particular to the field of performance optimization of data storage subsystems, in particular to an adaptive IO performance optimization method of data storage devices. Background technique [0002] With the rapid development of computer science and technology, computer systems have been widely used in various fields of social production and social life. The application of computer systems has played an irreplaceable role in the improvement of social productivity and people's quality of life. effect. Typical computer systems that are closely related to people's production and life include database systems, file server systems, and web server systems. The level of service quality of these computer systems is directly related to their performance level, high-performance computer systems will bring high-quality services, and vice versa. Therefore, in order to further ...

AI Technical Summary

Problems solved by technology

For example, the current IO request queue has 10 IO requests that are queued on a first-come-first-served basis. The data volume of the first request is 1GB, and the size of the next nine requests is less than 1MB. At this time, the first request Processing first, because of its large amount of data, the response time will be very long, and although the remaining 9 requests have been sent from the upper application system and entered the processing queue, they need to wait in the queue for a long time before they can be processed. The response time of the request is equal to the request waiting time and the request processing time. Although the request processing time of the subsequent 9 IO requests is relatively small due to the small amount of data, the long waiting time leads to a long response time.

In this case, the response time of all 10 IO requests is very long. As a result, the overall IO response speed of the external storage system is slow, the average response time of each IO request is long, and the IO performance is low.

[0006] The second point: the current IO performance optimization method lacks versatility and adaptability

The current idea of ​​IO performance optimization is to tailor a fixed optimization method for a certain type of application system at the device driver layer or storage device hardware layer. This fixed optimization method is only suitable for the IO load of the application system. If the upper layer If the application system changes, the characteristics of the IO load brought by the upper application system will also change, while the original device driver layer and storage device hardware layer will not change. At this time, the original optimization method cannot adapt to the new application. and new IO load

In this case, the original optimization method not only cannot play the role of IO performance optimization, but also often becomes an obstacle to the IO performance of the external storage system, and even seriously limits the IO performance of the computer system to the external storage device.

For example, if the original application system on the computer system is a video server, then the main feature of the IO load is the sequential reading of data. For this feature, the device driver adopts a sequential read-ahead strategy, that is, the external storage is read in advance according to the order of the storage address. The data on the device is for later use. This optimization method is appropriate for the video server, but if the upper-layer application system becomes a web server, the main feature of its load is random small amount of reading. At this time, there is no change The sequential read-ahead strategy in the device driver does not match this load characteristic, and cannot achieve the effect of IO performance optimization. On the contrary, the processor overhead and IO system overhead required for the implementation of this optimization strategy reduce the overall performance of the computer system.

Images