Method for realizing intelligent network card by combining FPGA with TCAM

Abstract

The invention discloses a method for realizing an intelligent network card by combining an FPGA (Field Programmable Gate Array) with a TCAM (Ternary Content Addressable Memory), which is characterized in that on the basis of unloading an OpenFlow flow table, a Megaflow flow table and a Microflow flow table are directly cancelled, so that the number of flow table items is reduced to a great extent, the TCAM is used for storing the OpenFlow table and corresponding execution actions, the work is stripped from the FPGA, and the operation efficiency is greatly improved. Therefore, the defects of the traditional design of the OVS full-unloading intelligent network card are overcome.

Description

technical field [0001] The invention belongs to the technical field of smart network cards, and more specifically relates to a method for realizing a smart network card by combining an FPGA with a TCAM. Background technique [0002] SDN is a stage product of the evolution of the network information age. OVS is an important application in the SDN field. Network operators can use OVS to realize private cloud and public cloud network access. OVS not only supports various network protocols, but also provides rich network functions with high flexibility and strong portability. However, since OVS is implemented at the software level and relies on the Linux kernel, the performance is low. As the number of virtual machines increases, each virtual machine has different requirements for network characteristics and performance. It is necessary to offload the network characteristics implemented by software to the underlying hardware and optimize the data flow path to improve network ...

AI Technical Summary

Problems solved by technology

The disadvantage of this design is that for unknown flows, the higher the correlation, the higher the hit rate, and the lower the update frequency of the flow table. On the contrary, the lower the correlation, the lower the hit rate, which will lead to a rapid increase of flow table entries , the flow entry aging speed is fast, and the system calls and flow table updates are frequent, which will cause the degradation of network communication quality and the reduction of offload capability to varying degrees.

Although the design structure of Megaflow flow table and Microflow flow table can be optimized to alleviate the problem of frequent update of flow table, it is always a temporary solution but not a permanent cure

Images