Novel cache file system framework

Abstract

The invention discloses a novel cache file system framework. The novel cache file system framework comprises an application program module and a bottom file system, wherein the application program module is used for acquiring data of a user service, and the bottom file system is used for storing the data of the user service; the system framework further comprises a high performance cache file system and a policy module, wherein the high performance cache file system is configured between the application program module and the bottom file system, and serves as a buffer layer of the bottom filesystem to store the data of the user service acquired by the application program module; the policy module is used for coordinating the data storage between the bottom file system and the high performance cache file system. The novel cache file system framework has the advantages that a cache file system is separately designed and developed, so that an original file system is not affected, the performance of the overall system can be quickly improved, and the novel cache file system framework is applied in a high performance scenario.

Description

technical field [0001] The invention relates to a computer storage file system, in particular to a novel cache file system architecture. Background technique [0002] File systems, especially distributed large-scale file systems, are currently the mainstream storage architecture. They are composed of general-purpose commercial servers and disks, provide large capacity, high performance, and high reliability, and are widely used in various industries. [0003] There are currently many popular distributed file systems, especially some open source file systems, because they use cheap general-purpose servers and the software is open source, so the cost performance is very high. But there was a serious problem: none of these filesystems performed well. There are roughly two reasons: one is because of open source software, the developers are scattered, and the structure and performance optimization are not enough; Low, which limits the application scenarios of this type of file ...

AI Technical Summary

Problems solved by technology

But there's a serious problem: none of these filesystems performs well

There are roughly two reasons: one is that the developers are scattered in the open source software, and the architecture and performance optimization are not enough; the other is that the general server is used, although the configuration is very high, but there is no targeted optimization and adaptation development, performance Low, which limits the application scenarios of this type of file system

[0004] However, this type of file system has rich interfaces, features, and functions, and has advantages in archiving and large-capacity cold storage. However, if it is insufficient in high IO scenarios, system stability, and performance mining for hardware, such as It is difficult to greatly improve the performance of NVMESSD

However, if you want to perform architecture optimization or code refactoring on such a file system, the workload is huge and the risks are huge.

Images