Design And Implementation Of Traffic Anomaly Detection Method In Power Integrated Data Network

In the power integrated data network,anomaly detection of business traffic is an important method to ensure network security and stability.However,with the continuous rapid development of the power integrated data network,the increase of various services carried has led to the explosive growth of network data traffic,which makes it difficult for existing anomaly detection methods to satisfy their demand.Therefore,it is necessary to study a new traffic anomaly detection method which is characterized by strong adaptability,high detection rate,low false positive rate and good timeliness.According to the demand of service traffic anomaly detection in power integrated data network,this paper proposes two methods of pre-screening with gray theory and KTLAD to improve original LOF anomaly detection algorithm.In the use of gray theory,gray theory is used to preliminarily screen the traffic to be detected and reduces the size of the set.In the KTLAD algorithm,the k-d tree is used to alleviate the problem of long time-consuming traversal of the original LOF algorithm.Both methods can reduce the cost of time consumption in traffic anomaly detection.Combined with the current development trend of big data technology,this paper selects spark platform to implement the proposed method for anomaly detection,which can better meet the computational requirement of the large amount of data and optimize its data computing strategy and storage way to improve the detection efficiency to satisfy the network security demand.In a word,this paper studies and implements a new traffic anomaly detection method for power integrated data network,solves the problems of weak adaptability,low detection rate,high false positive rate and poor timeliness in existing research,and further ensure the security and stability of the power integrated data network to improve the user's service quality.