As the network monitoring system topology of IDC security patrol robot is complex in structure with blind spots,its alarm system may cause the failure in accurately monitoring all security events or abnormalities.The design of IDC security patrol robot Network monitoring alarm system based on distributed Hadoop is proposed.The HDFS architecture and MapReduce task processing technology are introduced to design a Hadoop high-performance monitoring data processing platform for operation and maintenance cost reduction.The IDC security patrol robot Network monitoring system is designed on the platform to implement comprehensive patrol inspection on the data center.A threshold free alarm algorithm based on energy feature transformation is constructed to extract the three characteristics of energy,fluctuation and time in the monitoring item data.On this basis,the random forest algorithm is integrated into the model,automatically forming the patrol software design and trigger the alarm.The experimental results show that the proposed method has high alarm accuracy and efficiency and can effectively reduce IDC operation and maintenance costs.
维普
万方数据