A data processing method and chip for content router forwarding information base storage structure

Abstract

The invention discloses a content router FIB storage structure, which includes an internal storage unit and an external storage unit. The on-chip storage unit uses high-speed memory and deploys multiple index structures corresponding to the number of components with different name prefixes to realize fast indexing of name data based on the longest name prefix matching mechanism; the off-chip storage unit uses large-capacity low-speed memory and deploys multiple The FIB storage pool corresponding to the index structure to store the actual forwarding information. Among them, the on-chip index structure is implemented based on the neural network, and the storage efficiency is improved by learning the distribution of the index content in the memory. For possible misjudgments caused by the on-chip index structure, the chain address method is used in the off-chip FIB storage pool to deal with conflicts, that is, the data mapped to the same address is connected in the form of a linked list. The structure of the invention is suitable for the current memory hardware level, can effectively support the longest name prefix matching mechanism, and improve the retrieval speed of name data.

Description

technical field [0001] The invention belongs to the field of high-performance router storage structure design, and is particularly aimed at the structural design and data processing problems of the named data network content router FIB. Background technique [0002] With the continuous development of Internet and mobile communication technology, users are no longer satisfied with traditional point-to-point communication methods, but hope to widely distribute and share data information. However, the Internet communication model based on traditional TCP / IP technology has gradually exposed disadvantages such as exhaustion of IP address space, poor mobility, and low security, making it unable to meet users' requirements for low-latency data transmission and high-quality communication. For these reasons, NDN, as a new future network architecture, has attracted widespread attention once it was proposed. [0003] Compared with the traditional TCP / IP Internet, NDN uses a communicat...

AI Technical Summary

Problems solved by technology

However, the currently proposed FIB design scheme is difficult to balance storage consumption and name retrieval speed, and does not fully consider how to effectively support differentiated name retrieval and LNPM mechanism

Images