Method and system for classifying and storing data packets

Abstract

The invention discloses a data packet classification storage method and system. The method includes: obtaining the control information of the header of the data packet, matching and searching the corresponding rule address according to the control information; and storing the data packet in the data buffer , record the frame header information of the data packet in the address offset of the data buffer and the corresponding rule address in the data buffer storage table; when the data buffer is full, the external storage area takes out the data of the data buffer, and the file management system The physical block address of the external storage area is fed back, and the address block number of the external storage area of ​​the data and the offset value in the block are recorded in the index table through the record information of the data buffer storage table. The system includes a classification module, a buffer module and an external storage area. The invention can accelerate the execution speed of the data classification algorithm and improve the reusability of the hardware.

Description

technical field [0001] The invention relates to the field of network data transmission, in particular to a method and system for classifying and storing data packets. Background technique [0002] At present, the network demand is increasing with the development of the Internet, the services on the network tend to be complex and diverse, and the needs of customers are different. In order to adopt different services under different requirements, it is necessary to classify data packets, and store different types of data packets in different areas. [0003] Nowadays, a general-purpose central processing unit (CPU) or a dedicated integrated circuit chip is generally used to implement the packet classification function. However, the general-purpose CPU has problems such as insufficient I / O performance and high power consumption, while the special-purpose chip has problems such as insufficient I / O performance and high power consumption. The research and development cycle is long...

AI Technical Summary

Problems solved by technology

Among them, the linear search method is not efficient when faced with a huge rule set; the hierarchical tree method is easy to form a redundant structure, and it is difficult to establish and update; the recursive flow classification algorithm has a large storage overhead

The hardware method is superior to software in terms of speed and efficiency, but its implementation lacks flexibility

In the implementation method of ternary content addressable memory (Ternary Content Addressable Memory, TCAM), the support of special devices is needed, which makes the cost increase; the algorithm based on bit vector (BitVector, BV) is beneficial to realize on the hardware circuit, but in the circuit design Requires specialized hardware designers

Therefore, none of the above packet classification methods can meet the current complex and diverse network requirements.

Images