Lua coroutine-based program concurrent IO optimization method and system

Abstract

The invention provides a Lua coroutine-based program concurrent IO optimization method and system. The method comprises the following steps: when a service layer needs to execute IO operation, a coroutine running in the service layer circularly submits an IO monitoring request to an IO event through an intermediate interface; and after the IO event circularly waits for an IO handle corresponding to the IO monitoring request to be ready, the IO handle is called to execute a corresponding IO operation. According to the Lua coroutine-based program concurrent IO optimization method and system provided by the invention, a concurrent IO optimization scheme which simultaneously meets high concurrency, maintainability and good expandability can be realized.

Description

technical field [0001] The invention relates to the technical field of parallel processing, in particular to a Lua coroutine-based program concurrent IO optimization method and system. Background technique [0002] Today, with the development of the Internet and improved data, parallel processing of large amounts of data has become a daily basic scenario for servers. Based on this, the server's ability to handle concurrent IO has become an important indicator for judging business performance. [0003] For the requirements of concurrent IO, the traditional two types of solutions are the callback scheme based on asynchronous IO and the parallel scheme based on multi-process (thread). At present, some new solutions have been born in the industry, with their own advantages, such as gevent and go-routine. [0004] For callback schemes based on asynchronous IO (such as select and epoll), the form of callback cuts off the continuous function code context, which makes software deve...

AI Technical Summary

Problems solved by technology

[0006] For genvent, although concurrent IO is realized with the help of greenlet coroutines, there is no system-level thread switching overhead, and the development efficiency is extremely high, but the essence of greenlet coroutines is a supplementary implementation of python's native stack, and the memory usage is still large

[0007] For go-routine, the performance is higher, but some design ideas and characteristics of go (such as compiling, static linking, and built-in GC), which appear in large size and occupy more resources, determine its design adaptation is still General application layer scenarios for large high-end servers

Images