The Network Fault Management Based On The Rule In The IPv6 Network

With the development of our society, people all over the world go in to an age of technology and need more and more information. Since the 21th century, computer makes a great development, and in this case, the network naturally becomes a new way to transmit information. There are large numbers of communication on the internet, and the everyday life and work of people can't leave network. At present, the network technology of computer has a great influence on people's life. However, with the development of the network, the new problems also come out. In the process of the realization of the network, there are many mistakes and shortcoming. These faults and bugs go with scientists all the time. It makes the transmission velocity can't satisfy people's need. The computer network is not only consists one equipment, but includes many appliances and subnets. In this case, if any equipment in the network has something wrong may lead to the whole paralysis. These malfunctions will bring extremely loss. Therefore, the network fault management becomes a very important and urgent issue. Network fault management technology is from the computer network development, so it will be the cost of certain network resources. It is set in the network monitoring and management system construction, at any time on the network running status of each device monitoring and testing. About the early discovery of network faults, accurately locate the fault device is fault management system is facing major problems.In the management of network fault, the most fundamental and important part is the collection of fault information. There are many way to collect information. Consider the characteristic such as the high efficiency and standardization, we select the SNMP collection. At present, the network goes from IPv4 to IPv6. And the network fault management technology based on the IPv6 protocol is not very mature. Therefore, aim at the problem of IPv6, we advance the concrete plan. With the size of the network becoming bigger and bigger, and with the increasing complexity of network, the traditional management way can't competent for supervising the big and complex network. In this case, we utilize artificial intelligence technology to realize the intelligent management on network. Through the analysis of the existing technology and the merit summarization of old network management system, we advance a new kind of management model and then realize the system described by this model.The fundamental research work and innovation as follows:1. In order to manage the big size network based on the IPv6 environment and decrease the consumption of the resource of network and the burden of server in the management station, I advanced a new structure of management system which replaces the traditional system's flat structure. It is stratification structure which consists of management subsystem and fault collection subsystem. This kind of structure puts the network into many web pages. Each webpage is supervised by a fault collection subsystem. Fault collection subsystem filtrate, press, and analyze the information reported by every equipment. And then record the analysis result to the management station subsystem. Those information are deeply analyzed by management station subsystem.2. In order to resolve the problem of the two protocol coexistence, this model utilizes the popular IPv4/IPv6 double stack. The system itself is based on the IPv6 network environment. In the network we partly use IPv4 web page. We used the original SNMP protocol based on IPv4 to deal with date and then through SNMP Agent to get IPv6 MIB.3. This system used the newest device pattern, the program frame is based on the device mood which rely on the reversal principle and using connection in the crucial module. It made the fault analysis module and network port module based on relevant port device. In the future process of maintaining and updating system, we only need the module which to be replaced realize the port easily.4. In order to make the program more flexible and made the operation logic and code realization totally separated, the system's fault analysis and operation logic utilized the rule engine-Drools. In this case, after the system begin to run, if we need to filtrate alarm information, localize the fault and so on, we just need to update the operation logic which is used on rule engine. This kind of doing not only don't need to reboot the system, but also have no negative influence on the system which is running. Drools can use a statement to express business logic, you can also use non-XML in the local language rules, and Java code can also be very simple rules embedded into the file, so it is easy for us to learn and understand.5. The system I advance in this paper used a arithmetic called relevant rule which belongs to date mining in the module of fault analysis, filtration, localization. This kind of arithmetic can match rule engine very well which made the system have intelligent characteristics to some degree. Because the analysis and localization of fault play a vital role in the performance of system, it made the fault management model in this paper has very good characteristic.6. Because the system in this paper has a stratification structure, it made this structure of the system become very complex. In order to ensure the stability of system, I add the protection to each fault collection subsystem. The protection is a program which run in the fault collection system used the background way. It established a single environment for every fault collection subsystem to watch their running and consumption. Because the structure of the protection program is very easy, it can run very steadily. When the fault collection subsystem has a mistake or need to be rebooted, the protection program can accomplish these operations very well. 7. In the frame of the management station subsystem, I advanced a kind of device plan which is based on MVC. It considers the system's core function as Module to run on the server and consider the alternately part between the user and system as View module. This kind of doing can make the system support many servers'ending at the same time. And the server end can't be confined by C/S or B/S structure, and can even support this two patterns at the same time.8. At last, I summarized shortcoming and merits of this system and advanced three points may further study.In sum, I have done a quite comprehensive research to the network fault management system. I improved the shortcoming of original system and advanced a new system model and proposed a plan to crucial technology. This methods increased the efficiency and accuracy of prior similar system, and also made the system structure become more flexible and easily to maintain and update.