Research And Implementation Of Network Alarm Correlation Based On Fuzzy Association Rules

With the rapidly development of network and the popularity of Internet,almost every area of people's lives and social economy are increasingly rely on the Internet.In this way,the security of computer network will become more and more important,as the scale of network is expanding-and the network structure is more and more complex.However,the hardware and software service in network will produce a lot of network alarm which will make the operation of network in an abnormal state.The alarm in network is in a huge amount as the equipments are correlated,even will lead to alarm storm.So we should pay more attention on network alarm.However,there are some relationships between each alarm and if we could use some methods to discover the hidden relationship of these alarm,we will help network manager locate the fault and make the network into normal state.And the ability of whether the network could quickly recover from the bad state is a measure of the performance of a network management.To solve the above problems,this article use association rules mining based on fuzzy set to explore the relationship of network alarm.In this paper,firstly,some knowledge and technology of traditional data mining is introduced and one of the famous association rule algorithm,Apriori algorithm is introduced.Secohdly,we make an analysis of the limitation of traditional association rules and introduce a more suitable association rules mining algorithm.Thirdly,we try to use the fuzzy association rules mining algorithm to discover the hidden relationship between network alarms.And make a experimental analysis to prove it.