NETWORK ABNORMAL STATE DETECTION DEVICE USING HMM(HIDDEN MARKOV MODEL) AND METHOD THEREOF

摘要

A device and a method for detecting a probabilistic abnormal state of network using Hidden Markov Model are provided to enhance the accuracy of determining a network error by HMM(Hidden Markov Model). An error symptom extracting unit(410) analyzes a network traffic and outputs the network traffic characteristic. An error symptom determining unit(420) uses a network traffic characteristic as an input value, and outputs each state of network by a probability value. The error symptom extracting unit comprises a traffic input unit(412), a packet processing unit(414), a log recording unit(416), and a network traffic characteristic outputting unit(418). The packet processing unit processes the inputted network traffic as packet-by-packet to check the status of the network traffic.