Fast indexing method

Abstract

The method segments all indexed address in same type, and establishes tree root table for address of prefix segment of each segment, and establishes index table for addresses in each segment. The method includes steps: according to the said segment mode to segment index address to be indexed; determining current index table; if determining current table to be indexed is as index table in first segment, then using first segment address as keyword to carry out indexing in index table of first segment, ending flow; if determining current table to be indexed is as tree root table, then using address of prefix segment corresponding to tree root table as keyword looks up table item matched to the address of prefix segment in tree root table; if finding it out, then carrying out indexing by using the matched pointer of table item as entry, and using address of corresponding segment as keyword; ending this flow; if not finding it out, then determining current index table again.

Description

technical field [0001] The invention relates to an indexing technology in the field of communication computers, in particular to a fast indexing method. Background technique [0002] The rapidly developing Internet will no longer be a network that only connects computers, but will develop into an information and communication infrastructure similar to the telephone network and cable TV network. Therefore, the Internet Protocol (IP, ie IPv4) being used is already incompetent. In order to meet the growing needs of the network scale, the Internet Engineering Department (IETF) has designed the next generation Internet Protocol IPv6. Compared with IPv4, IPv6 has undergone a huge change in the address format, and the address length has changed from the original 32 bits (bit) to 128 bits. [0003] After several generations of development, IP routing technology has developed from the initial centralized software routing technology to the current distributed routing technology with ...

AI Technical Summary

Problems solved by technology

[0010] The T-tree routing technology of IPv4 mentioned above is a hierarchical indexing method. If IPv6 uses the T-tree routing technology of IPv4, since the address length of IPv6 is four times that of IPv4, there will be obvious disadvantages: IPv6 addresses are 8bit When building an index for the first level, it is necessary to build a 16-level index, which causes the forwarding speed to drop to 25% of that of IPv4, and requires 4 times the hardware resources

Therefore, the routing index speed decreases, the memory consumption is large, and the hardware resources required are also large.

[0018] 2) When using HASH technology for IPv6 routing index, if a HASH index table with a small address space is established, due to the long length of IPv6 addresses, one address in the index table corresponds to many host addresses, so correct routing cannot be guaranteed; The HASH index table with a large address space consumes a lot of hardware resources

In the HASH routing index, the index keywords must be of the same length, so network segment routing cannot be supported

[0019] 3) When using CAM auxiliary device for IPv6 routing index, due to the limited number of routing indexes supported by CAM, it cannot meet the needs of IPv6 routing index; if the number of CAMs is increased to increase the number of routing indexes, the hardware cost will be greatly increased

Images