Linux 是目前广泛用于路由设备中的操作系统,而流量管理是这种网络操作系统的一个重要功能.研究了 Linux 系统的流量管理机制,发现当前 Linux 系统所采用的在网络接口的出口实现的基于网络包调度的流量管理机制缺乏对于网络带宽在系统...Linux 是目前广泛用于路由设备中的操作系统,而流量管理是这种网络操作系统的一个重要功能.研究了 Linux 系统的流量管理机制,发现当前 Linux 系统所采用的在网络接口的出口实现的基于网络包调度的流量管理机制缺乏对于网络带宽在系统范围的全局性的管理,也缺乏对于输入流的网络处理的有效管理和调度,从而造成不必要的 CPU 资源的浪费.为解决这一问题,提出了一种新颖的网络带宽管理机制,它通过调度网络协议处理所占用的 CPU 时间来实现带宽管理.这种新的机制将网络带宽的管理和调度从网络接口的出口点转移到处理接收到的网络包的系统软件中断处理程序中,从而克服了原来的流量管理机制的缺点.通过系统实现和对比实验,发现该机制本身给系统带来的负载小于 Linux 原来的流量管理机制,同时可以提供更好的流量隔离,并且能够有效地节省用于网络处理的 CPU 资源.展开更多
针对信道资源有限的高级数据链路控制(High Level Data Link Control,HDLC)通信,提出了一种零开销的带内管控技术。该技术采用输入输出异步缓存、高频管控帧插入、高可靠管控帧提取的方式,构建带内管控数据通道,实现了对HDLC帧数据的零...针对信道资源有限的高级数据链路控制(High Level Data Link Control,HDLC)通信,提出了一种零开销的带内管控技术。该技术采用输入输出异步缓存、高频管控帧插入、高可靠管控帧提取的方式,构建带内管控数据通道,实现了对HDLC帧数据的零开销管控。该技术可有效解决HDLC帧端到端满带宽通信情况下,零开销带内管控的设计难点。最后,给出了基于现场可编辑逻辑门阵列(Field Programmable Gate Array,FPGA)的实现方案和功能仿真验证结果。展开更多
The evolution of computer networks has experienced several major steps, and research focus of each step has been kept changing and evolving, from ARPANET to OSI/RM, then HSN （high speed network） and HPN （high perfo...The evolution of computer networks has experienced several major steps, and research focus of each step has been kept changing and evolving, from ARPANET to OSI/RM, then HSN （high speed network） and HPN （high performance network）. During the evolution, computer networks represented by Internet have made great progress and gained unprecedented success. However, with the appearance and intensification of tussle, along with the three difficult problems （service customizing, resource control and user management） of modern network, it is found that traditional Internet and its architecture no longer meet the requirements of next generation network. Therefore, it is the next generation network that current Internet must evolve to. With the mindset of achieving valuable guidance for research on next generation network, this paper firstly analyzes some dilemmas facing current Internet and its architecture, and then surveys some recent influential research work and progresses in computer networks and related areas, including new generation network architecture, network resource control technologies, network management and security, distributed computing and middleware, wireless/mobile network, new generation network services and applications, and foundational theories on network modeling. Finally, this paper concludes that within the research on next generation network, more attention should be paid to the high availability network and corresponding architecture, key theories and supporting technologies.展开更多
文摘Linux 是目前广泛用于路由设备中的操作系统,而流量管理是这种网络操作系统的一个重要功能.研究了 Linux 系统的流量管理机制,发现当前 Linux 系统所采用的在网络接口的出口实现的基于网络包调度的流量管理机制缺乏对于网络带宽在系统范围的全局性的管理,也缺乏对于输入流的网络处理的有效管理和调度,从而造成不必要的 CPU 资源的浪费.为解决这一问题,提出了一种新颖的网络带宽管理机制,它通过调度网络协议处理所占用的 CPU 时间来实现带宽管理.这种新的机制将网络带宽的管理和调度从网络接口的出口点转移到处理接收到的网络包的系统软件中断处理程序中,从而克服了原来的流量管理机制的缺点.通过系统实现和对比实验,发现该机制本身给系统带来的负载小于 Linux 原来的流量管理机制,同时可以提供更好的流量隔离,并且能够有效地节省用于网络处理的 CPU 资源.
文摘针对信道资源有限的高级数据链路控制(High Level Data Link Control,HDLC)通信,提出了一种零开销的带内管控技术。该技术采用输入输出异步缓存、高频管控帧插入、高可靠管控帧提取的方式,构建带内管控数据通道,实现了对HDLC帧数据的零开销管控。该技术可有效解决HDLC帧端到端满带宽通信情况下,零开销带内管控的设计难点。最后,给出了基于现场可编辑逻辑门阵列(Field Programmable Gate Array,FPGA)的实现方案和功能仿真验证结果。
基金supported in part by the National Natural Science Foundation of China under Grants No.90604003 and No.90604004by the National Grand Fundamental Research 973 Program of China under Grant No.2003CB314801.
文摘The evolution of computer networks has experienced several major steps, and research focus of each step has been kept changing and evolving, from ARPANET to OSI/RM, then HSN （high speed network） and HPN （high performance network）. During the evolution, computer networks represented by Internet have made great progress and gained unprecedented success. However, with the appearance and intensification of tussle, along with the three difficult problems （service customizing, resource control and user management） of modern network, it is found that traditional Internet and its architecture no longer meet the requirements of next generation network. Therefore, it is the next generation network that current Internet must evolve to. With the mindset of achieving valuable guidance for research on next generation network, this paper firstly analyzes some dilemmas facing current Internet and its architecture, and then surveys some recent influential research work and progresses in computer networks and related areas, including new generation network architecture, network resource control technologies, network management and security, distributed computing and middleware, wireless/mobile network, new generation network services and applications, and foundational theories on network modeling. Finally, this paper concludes that within the research on next generation network, more attention should be paid to the high availability network and corresponding architecture, key theories and supporting technologies.
