Analysis And Research On Several Basic Problems In Network Traffic Monitoring Equipment

With the development of network technolody, network application and network scope. The Internet has become the world people's necessary tools and important information dissemination channels in work, live and play. For many people in their daily lives have become accustomed to listening to music using the Internet, view news, instant messaging with others, view network video and network television, using search engines to find information, e-mail, play online games, in the blog, microblogging and personal space to express their articles and comments, chat and forums conducted through online shopping and more. Today's Internet provides a variety of businesses and services, so people can transmission of text, audio, images, animation, video and other types of data via the internet. People in the enjoyment of the Internet has brought many advantages and benefits, but also more and more be influenced by network congestion, network security, network troubles and bad information, and other hazards.With the rapid development of Internet technology, network users, network services, network traffic is the explosive growth. Rich variety of network services not only enrich the network, but also give a great challenges to network management and monitoring. In this background, the face of increasingly complex network of online and increasing network traffic, network managers must spend more time and effort to understand the operation status of network devices to maintain the normal operation of a network. Network traffic management and monitoring provided that the first of the network traffic identification and classification, as there is a wide variety of network services, in particular the emergence of P2P services, network traffic, the explosive growth, and with the A variety of services purposes to avoid the increase of the phenomenon of network monitoring, network services continue to adopt new encryption method, an increase of network'traffic classification very difficult.How to better the traffic on the network classification, statistics and monitoring, is the key problem of the network traffic monitoring technology. For the current Internet services development and the status of the network management requirements, this paper forcus on the network traffic monitoring equipment, mainly from the perspective of achieving the classification of network traffic, traffic statistics and analysis, and other aspects of flow control algorithms research. Forcus on the network traffic monitoring technology, with the purpose of proposing related algorithms of this technology, this paper mainly works on the network traffic classification, traffic monitoring system infrastructure, flow compression and statistical records, flow control algorithms, and domain name blocking. Main contents are:1) Hardware-based flow classification model. This paper presents a TCAM-based hardware string matching algorithms, and proposed an algorithm based on FPGA and Bloom Filter string matching algorithms. Combination of DPI and the DFI technology research proposed hardware-based flow classification model.2) Flow statistics and analysis based on DPI technology. In this paper, DPI-based flow classification and statistical techniques were studied, a DPI system infrastructure. Proposed hash-based compression algorithm of long flow label and Identifying elephant flows algorithm, which all can be use in DPI system.3) Flow Control Algorithm Analysis and Improvement. In this paper, basing on the research and analysis of the token bucket algorithm, a new token bucket cascade of shared user bandwidth, security algorithms is Propsed. And at the same time, we compare and study the different between the packet loss flow control algorithm and the packet disturb flow control algorithm. 4) Hardware-based algorithm for host domain name blocking. This paper presents an algorithm of blocking illegal websites by internet host domain basing on FPGA and Bloom Filter algorithm. At the same time, for high-speed 40Gbps DPI devices, we proposed a TCAM-based host domain blocking algorithm.