Based Expert System For Network Fault Detection Method And To Achieve

Ever since the eighties of last century, with the rapid development, network has been used in financial,education and more areas. With the indepth application of network, the effciency of those fields has been increased very much. On the other hand, the application of network has deepened the dependence to network. Once the network were in trouble, it will cause a great of losses. For the reason, an effective detection method for network trouble will be very important.Because of the complexity of network protocol and environment, it has made an effective method design for network trouble become a tough problem. This paper aims to design and implement a set of methods to detect network trouble automatically. By taking the TCP/IP model as a model, against the link layer and network layer, based on indepth analysis about the function of layers and important protocols, we design and implement 23 kinds of detection methods. For the automanization of those network trouble detection methods, we consult the ideal of expert system, and make use of CLIPS expert system tools to implement them. The main work and result can be summarized as follows:(1) The total design of network trouble detection methods design and implementation. Take the protocols of link layer and network layer in TCP/IP model as a clue, we analysis the design requirment about network trouble detection, and decide the total architecture. Because of qucik speed and easy to port, we choose CLIPS as expert system tools to implement the mothods we designed.(2) The design of network trouble detection method. According to the sequence from down to up about TCP/IP model, we take design for link layer and network layer trouble detection. About link layer, we concern error frame and ARP protocol and design 7 kinds of network trouble detection methods. About network layer, we designed 16 kinds of trouble detection methods around IP and ICMP protocol. Because of the complexity about IP fragmentation and reassemble, it has induced the detection of IP fragmentation become a difficult point. Here we design a method which detects IP fragmentation missing by virtual IP fragmentation reassemble. This method is not only effective, but also can avoid cumbersome data conduct for the IP fragmentation.(3) The implementation of network trouble detection method which is based on expert system. Here we introduce the basic thory of expert system, analyse the work pattern of CLIPS expert system tools, and choose IP fragmentation missing detection etc. 3 kinds of typical network trouble detection methods to explain how to use CLIPS to implement network trouble detection.