The Design And Implementation Of IPv6NetStream Software System

NetStream technology developed by H3C is used to analyse network traffic based on flow. The goals of IPv6NetStream software system is to analyse network traffic in IPv6network according to user’s configuration, to provide technical support for a series of applications,for example, network billing,planning analysis, network monitoring, application monitoring and analysis.In the requirement phase of IPv6NetStream software system, firstly, understand comprehensively users’nedds and carry out a detailed analysis and research, and then use a data flow diagram to describe the NetStream business processes. The functions of this system are to support IPv6packet statistics, support for streaming aging, support for streaming aggregation, support V9format output, support for virtualization, support for distribution, support sending flow control, support on-demand start, to realize IPv6NetStream configuration commands and display commands.In the design phase of IPv6NetStream software system, we take a structured design method named Layer by layer from top to bottom to analyse this system. IPv6NetStream software system is divided into two main sections:user mode and kernel mode. IPv6NetStream in user mode has a command-line plugin NetStream CLIPlugin, NetStream LIB to offer an external interface and NetStream Daemon. NetStream CLIPlugin completes the definition,registration and analysis of the IPv6NetStream configuration commands. NetSteam LIB provides external interfaces to encapsulate configuration commands according to designed patterns and to send messages to the NetStream Daemon. The tasks of NetStream Daemon are to initialize all resource configuration, analyse messages from NetStream CLIPlugin, save configuration data in DBM, handle a variety of system events, respond to timers events, restore configuration and synchronize configuration. IPv6NetStream in kernel mode is responsible for implementing the main function of system:the flow statistics, flow management, flow-aging, flow aggregation, outputed flow control, template management, sent templates, thread management, handle ip address event handle MDC event, maintain configuration data and so on.For this system, its unique is that statistical packet base on template so that it has a lot of flexibility and scalability, the system only uses one set of code to support a huge number of products whose structure is very different. However, there are many aspects can be further improved. We can make further efforts to enhance the efficiency of network traffic classification and statistics. We can do our best to reduce the ipv6packet loss rate when there are plenty of ipv6packets suddenly in the computer networks. In the future, we can take these aspects as very important study directions to improve system performance.