Research On Technologies Of Data Collection And Management For SNMP Based Network Operation State Monitoring

Under the tendency towards large scale and diversification of the network, it has become a hybrid network with multi-vendor and a large number of devices coexisting and complex and changeable technologies. Under such conditions, the security mangagement of the network faces severe challenges. Especially, in the SNMP based network monitoring, it is hard to monitor all the network devices for data collection. So it is needed to select monitors optimally so as to realize effective network monitoring. Besides, in the work of the monitoring data management, a lot of specific operations, complicated operational procedures, and frequent repeated patterns, make the the work of the data management low efficiency and the operations very cumbersome. Using data configuration template can preserve the specific operations, operational procedures, and the repeated patterns and reuse them such that simple and efficient data management operations can be realilzed.But, in this thesis, we find if it is only needed to cover a portion of traffic with a given cover ratio, there can be multiple solutions satisfying the optimization requirements according to the former optimal monitor selection methods. And the data in the configuration template cannot be reused sufficiently because of its data acquisition method using the static and unitary index. Consequently, based on the theory of the high variability of the network node degrees and the knowledge management, in the thesis we propose new mentalities and solutions. The main work is outlined as follows:1. An algorithm for selecting effective network traffic monitors based on cover ratio is designed. This thesis proposes a solution of pratical value for selecting monitors based on the critical node identification in the network, which is a cost-effect trade-off. It realizes the best monitor selection for a given traffic cover ratio. The simulation experiment verifies the vadility of the algorithm. From the perspective of the representative and the importance, the selected nodes not only have good coverage effect, but also have strong segmentation ability.2. A mechanism is designed to implement the ontology based data configuration template. To solve the problems in the existing data configuration template, the thesis raises a way of thinking for targeted data reuse. Recogonizing the data manipulation and the existing data configuration template as knowledge, and using the ontology to construct the management data knowledge base, we propose a complete set of mechanisms for the effective acquisition, discovery, and reuse of the management knowledge, which satisfies the differentiation and customization of the demand in the data management manipulation.3. The domain ontology applicable to the SNMP based monitoring is designed and developed for the data management. According to the application requirements of the data reuse, on the basis of the indepth study of the management data knowledge of the SNMP based monitoring, we develop the network monitoring data management ontology using the Protégé. It lays the foundation of the knowledge acquisition, discovery, and reuse in the monitoring data management. 4. A prototype system for task feature oriented monitoring data management is designed and implemented. Based on architecture of Struts, using the semantic web application framework Jena and the querying and reasoning engine Pellet, we develop a prototype system with function of feature association and template automatic generation. It verifies the feasibility and validity of the theory of the ontology based data configuration template proposed in the thesis.By the above work, the research on the monitoring data collection and management is strengthened and complemented, and the design of the optimal network monitoring solution and the data management work wth the simplicity and high efficiency.