High-performance fluid control method based on linux kernel

Abstract

The invention relates to a high-performance fluid control method based on linux kernel. The high-performance fluid control method comprises the following steps of: forming a mark field in sk_buff of the Linux kernel;, adding an index field into the mark field, wherein the Mark field is used for performing fluid control on a bandwidth object, and index is an unsigned integer, and dividing the index field into two parts, wherein high 16-bit is used for performing service fluid control, and low 16-bit is used for performing single ip fluid control; establishing a data packet rank; controlling whether the data packet is out of the rank through using the fluid control method, namely controlling the data packet according to the service fluid control method, if the data packet is out of the rank, sequentially controlling the data packet through using a bandwidth object fluid control method and a single IP fluid control method, and if the data packet is not out of the rank in any one of the three fluid control methods, ranking again; and finally, selecting the data packet out of the rank through using a packet transmission scheduling method and transmitting the data packet. The performance of the single ip fluid control is higher than the performance of a module in the conventional linux, the waste of out-of-rank time is avoided, and the performance is improved.

Description

Technical field: [0001] The invention relates to the shaping and control of flow by gateway products, in particular to a high-performance flow control method based on the linux kernel. Background technique: [0002] The inherent flow control strategy of Linux is divided into two types: with queue and without queue, which can meet the simple needs of some users for flow control, but there are still many deficiencies in many scenarios. [0003] Linux inherent technical flaws: [0004] 1. In these technologies, either the user controls certain network segments, or controls the flow of certain hosts. It is impossible to control certain network segments at the same time and cannot exceed a certain flow rate. At the same time, the service and each host cannot exceed a certain flow rate. [0005] 2. The Htb algorithm is currently the most used flow control algorithm in Linux, which can solve some problems very well. However, when the user needs to control the flow of each host i...

AI Technical Summary

Problems solved by technology

[0004]1. In these technologies, either the user controls certain network segments, or controls the traffic of certain hosts. At the same time, the service and each host cannot exceed a certain amount of traffic

[0005]2. The Htb algorithm is currently the most used flow control algorithm in Linux, which can solve some problems very well, but when the user needs to control each host in the network segment When the flow rate cannot exceed a certain amount, and the number of hosts is large, this consumes a lot of performance for the htb algorithm and cannot achieve real practicability

Images