TCP channel management system and method suitable for high-concurrency scenes

Abstract

The invention discloses a TCP channel management method suitable for high-concurrency scenes. The method comprises an access gateway, message middleware, a data analysis service and a TSP business service, wherein the access gateway is connected with a plurality of clients; the access gateway s connected with the message middleware; the message middleware is connected with each of the data analysis service and the TSP business service; and the access gateway is used for providing services such as client access, link validity verification and overtime management. According to the method, an epoll asynchronous communication model is adopted, a task pool, link table management and a link detector are imported, and a task queue pool and a multi-thread technology are combined, and a communication link is decoupled. The method is suitable for high-concurrency application scene characteristics and network communication technology background, and sufficiently considers the advantages and disadvantages of various technologies so as to adopt the advantages and avoid the disadvantages.

Description

technical field [0001] The present invention relates to the field of telecommunications, in particular to a TCP channel management system and method applicable to high concurrency scenarios. Background technique [0002] When doing network programming, there are four common calling methods: synchronous (Sync) / asynchronous (Async), blocking (Block) / non-blocking (Unblock). The difference between synchronous IO and asynchronous IO lies in whether the process is blocked during data access; blocking The difference between IO and non-blocking IO is whether the call of the application program returns immediately; select and epoll are solutions for multiple I / O multiplexing provided by the Linux kernel, and select essentially stores the data of the fd flag by setting or checking Disadvantage 1 is that the number of fds that can be monitored by a single process is limited, that is, the size of the listening port is limited; Disadvantage 2 is that the socket scans linearly, that is, t...

AI Technical Summary

Problems solved by technology

[0002] When doing network programming, there are four common calling methods: synchronous (Sync) / asynchronous (Async), blocking (Block) / non-blocking (Unblock). The difference between synchronous IO and asynchronous IO lies in whether the process is blocked during data access; blocking The difference between IO and non-blocking IO is whether the call of the application program returns immediately; select and epoll are solutions for multiple I / O multiplexing provided by the Linux kernel, and select essentially stores the data of the fd flag by setting or checking Disadvantage 1 is that the number of fds that can be monitored by a single process is limited, that is, the size of the listening port is limited; Disadvantage 2 is that the socket scans linearly, that is, the method of polling is used, and the efficiency is low; The implementation of epoll is based on the callback function on each fd of the kernel. Only active sockets will actively call the callback, so when there are fewer active sockets, using epoll does not have the performance problem of linear decline of select, but all When the socket is very active, there may be performance problems

Images