Storing, searching method and apparatus for data packet

Abstract

The invention discloses a data message storing method, a data message searching method and devices thereof. The data message storing method comprises performing a hashing process to data messages according to a linear hashing function and obtaining the first index information of the data messages; judging whether the depth of the current hash bucket is smaller than a set standard value of the bucket depth or not according to the first index information of the data messages, if so, the data messages being storied in the current hash bucket, if not, judging whether the depth of a related hash bucket of the current hash bucket as the related hash bucket rule defined is smaller than the set standard value of the bucket depth or not, if so, performing a quadratic hashing to the data messages according to the designed quadratic hashing function and setting a splitting label for the current hash bucket. The invention also provides a data message storing device, a data message searching method and a data message searching device. According to the invention linear searching times in a hash bucket is reduced and searching efficiency of data messages is improved.

Description

technical field [0001] The invention relates to switching technology, in particular to a method and device for storing and searching data messages. Background technique [0002] Usually, when the switch processes data packets, it needs to send a part of data packets that are more complicated to be processed to the CPU for corresponding processing. Before sending the data packets to the CPU, these Data messages are stored by category. When sending a message to the CPU later, it needs to be searched in the HASH list to determine whether the message has been stored. The essence of the HASH search algorithm is to use the HASH function to calculate the corresponding output value through the input data message, thus improving the search efficiency of message data. However, due to the existence of HASH conflict chains, it is necessary to perform a linear search in the conflict chain. The linear search reduces the overall efficiency of the HASH algorithm, which in turn affects the ...

AI Technical Summary

Problems solved by technology

However, due to the existence of HASH conflict chains, it is necessary to perform a linear search in the conflict chain. The linear search reduces the overall efficiency of the HASH algorithm, which in turn affects the packet processing capability of the CPU. Moreover, as the length of the HASH conflict chain increases, its impact will decrease. bigger

[0003] There are many ways to improve the efficiency of HASH search in the prior art, one of which uses the idea of ​​HASH algorithm to establish the corresponding relationship between the message number and the message processing function, and uses the technology of reference counting sorting to sort the data with high query frequency The text is stored in front of the conflict chain, but it does not consider how to solve the problem of HASH conflicts. Some long HASH conflict chains still need to be linearly searched, so the search efficiency can only be improved to a certain extent.

Another method uses the HASH linked list to store user equipment information, uses the open address method and transfers the conflicting hash data message to the binary tree structure to solve the HASH conflict problem, this method can effectively solve the HASH conflict, but The implementation is more complicated, and the application scenarios are relatively narrow

Images