A method, system and device for stack layer optimization judgment

Abstract

The present application discloses a method for stack layer optimization judgment, comprising: receiving input test information; determining a call command, a stack layer to be tested, and a call layer according to the test information; virtually generating a first call in the call layer according to the call command result, and obtain the first performance data according to the first call result; virtually generate the second call result according to the call command in the stack layer to be tested, and return the second call result to the call layer, and then obtain according to the second call result second performance data; determining whether the stack layer to be tested needs to be optimized according to the first performance data and the second performance data. This application performs IO tests on each stack layer in the IO stack by returning virtual data, and then analyzes the impact of each stack layer on the system IO, so as to point out the direction for system IO performance optimization and narrow the scope of IO tuning. At the same time, the present application also provides a system, equipment and computer-readable storage medium for stack layer optimization judgment, which have the above-mentioned beneficial effects.

Description

technical field [0001] The present application relates to the field of IO performance, and in particular to a method, system, device and computer-readable storage medium for stack optimization judgment. Background technique [0002] Computer system IO performance tuning is a problem that IT equipment manufacturers and network companies cannot avoid. From simple parameter tuning of general computer system IO performance to changing the system-related IO architecture to adapt to relevant scenarios, the purpose is to hope that the system's IO performance can meet the user's IO performance requirements. [0003] There are two main reasons why IO performance optimization is a difficult problem for IT equipment manufacturers and network companies: [0004] 1) The IO path of the computer system is very long. Even if network access is not considered, the IO path of the desktop linux system may include the following layers: user application → system call → virtual file system → file...

AI Technical Summary

Problems solved by technology

A system that satisfies a certain scenario may not be applicable at all when switched to another scenario, and serious data loss may occur

Therefore, if system tuning wants to achieve satisfactory tuning results, it is necessary to have a good understanding of the performance of the entire stack layer of the system, and only using relevant tools for test analysis may not be able to obtain real data, and thus cannot correctly analyze system performance bottlenecks

Images