Method for optimizing TCP (Transmission Control Protocol) connection management by adopting static linked list to structure TCP node pool

Abstract

The invention provides a method for optimizing TCP (Transmission Control Protocol) connection management by adopting a static linked list to structure a TCP node pool. A connection management structure comprises a TCP concurrent connection hash table and a TCP node pool; when a system is initialized, all the TCP nodes required by a memory are applied once and are serially connected in a static linked list to form a TCP node pool; when one TCP connection is established, one node only needs to be taken out from the node pool and hung into the hash table; when one TCP connection is deleted, the TCP node only needs to be taken down from the hash table and placed back to the node pool; and thus, the memory application and release operations during frequent establishment and closure of the TCP connections are avoided in the system operation process, and the efficiency of the TCP connection management system is increased.

Description

technical field [0001] The invention relates to the field of network data processing, in particular to a method for optimizing TCP connection management by constructing a TCP node pool by using a static linked list. Background technique [0002] TCP connection management is an important part of the TCP protocol stack. It is necessary to use a concurrent TCP connection table to save each current TCP connection. When a new TCP connection is established, a new entry needs to be created in the TCP connection table. When a When the TCP connection is closed, the connection entry needs to be deleted from the concurrent connection table. [0003] On a high-speed network monitoring system, it is often necessary to manage millions to tens of millions of concurrent connections. The establishment and deletion of TCP connections are very frequent, and the number of new and closed connections per second is sometimes as high as hundreds of thousands. Connections must apply for and release...

AI Technical Summary

Problems solved by technology

[0003] On a high-speed network monitoring system, it is often necessary to manage millions to tens of millions of concurrent connections. The establishment and deletion of TCP connections are very frequent, and the number of new and closed connections per second is sometimes as high as hundreds of thousands. Connections must apply for and release the TCP node structure, that is to say, frequent application and release of small memory will significantly reduce the efficiency of the entire system

Images