| With the rapid development of data collection and storage technology, the field of network management systems have gradually accumulated logs with large data, affecting the efficiency of network management and system stability directly. Achieving the correlation analysis is an important and basic requirement of network management system. Combined with other technologies such as artificial intelligence or database, rules engine technology is proposed to solve this problem.Derived from the reasoning engine, Rules engine employs pattern matching method, and detaches the business rules and the application code. Based on the introduction of the related research background and basic principle, this dissertation mainly includes the following important contents:1. In this dissertation, I introduce some net log management technology, and analyze the network management log analysis system structure based on rules engine. Then I give a detailed expatiation on the kernel Rete algorithm of the rule engine, and demonstrate the algorithm process.2. I study and analyse the rete algorithm from the performance efficiency. On this basis, I also improve the algorithm from the following aspects:facts change in the memory, storage space management and the hash table application.Then I compared the improved algorithms with the original ones in the experiment; the result showed that the time performance has improved a lot, without sacrificing larger space.3. I design the alarm log the correlation analysis system of the network failure management system, and develop the business and rules module, then illustrate log collection and formattern processing method, and design the interface and data structure, the acquisition of the correlated rules and rules library storage. Then I employ Drools rules engine tecnology to realize the whole process of log correlation analysis in the Eclipse platform with Java language; Finally we have perfected the main core functions. |
PDF Full Text Request