|国家科技期刊平台
首页|期刊导航|现代电子技术|基于FPGA的ICN名字解析缓存加速系统

基于FPGA的ICN名字解析缓存加速系统OACSTPCD

Design of ICN name resolution acceleration system based on FPGA

中文摘要英文摘要

名字解析服务是ICN网络中获取信息内容的关键.为提升名字解析服务的效率,降低名字解析系统的负载,设计并提出一种基于FPGA的ICN名字解析缓存加速系统.该系统在网络节点的数据面上利用FPGA卸载名字解析功能,实现标识到网络地址的映射存储,处理需要查找网络地址的转发数据包,同时,该系统支持控制面动态配置映射关系,保证映射存储的更新.实验测试结果表明,该系统在10 Gb/s的测试流量下,对小包的转发率达到4.6 Mpps,并且请求延时低于1.5 μs,相较于传统软件实现的键值存储系统处理速度提高了27倍,与使用DPDK内核旁路进行优化的软件相比处理速度提高了4倍.

Name resolution service is the key to retrieving information content in ICN(information-centric networking).Therefore,an FPGA-based ICN name resolution acceleration system is designed and proposed to enhance the efficiency of name resolution services and reduce the load on name resolution systems.This system offloads the name resolution function to FPGA on the data plane of network nodes,achieves the mapping storage of identifier-to-network address and processes forwarding data packets that require to find the network address.Additionally,the system supports dynamic configuration of mapping relationships on the control plane,and ensures the update of mapping storage.Experimental test results show that the proposed system achieves a forwarding rate of 4.6 Mpps for small packets when the test traffic is 10 Gb/s,with a request latency below 1.5 μs.In comparison with the traditional software-based key-value storage systems,the processing speed of the proposed system is increased by 27 times,and it is 4 times faster than the processing speed of software optimized with DPDK kernel bypass.

李雪彤;陈晓;宋磊

中国科学院声学研究所 国家网络新媒体工程技术研究中心,北京 100190||中国科学院大学,北京 100049

电子信息工程

ICNFPGA名字解析键值存储SDN哈希表硬件加速流水线

ICNFPGAname resolutionkey-value storageSDNHash tablehardware accelerationpipeline

《现代电子技术》 2024 (007)

109-114 / 6

先导专项课题(XDA031050100)

10.16652/j.issn.1004-373x.2024.07.019

评论