Hidden Markov Model Based Traffic Identification Technology of Satellite Communication Network
Jiaying Li, Liujing Hu, Xianghui Hu, Zhao Liu, Liyuan Li, Zhixin Liu · 2022 IEEE International Conference on Unmanned Systems (ICUS) · 2022
In the satellite communication network, it is significant to realize real-time identification of specific application traffic for ensuring the Quality of Service (QoS). A traffic identification technology based on Hidden Markov model (HMM) is proposed, which establishes a specific application traffic identification model that relies on the characteristics of packet arrival order, arrival time gap and packet size. The sample traffic packets map to HMM states, and some discrete random variables are used to represent the fixed state characteristics of HMM. This technology can shorten the model establishment time, and improve the accuracy and real-time performance of identifying application traffic. This paper proposes an optimal state count selection algorithm and designs the Network Application Identification Architecture (SNAIA). Eventually, the testing envirnoment is built in the virtual satellite communication network simulation system. SNAIA is used to identify a variety of application traffic. Results show that SNAIA can quickly and accurately identify the packet traffic generated by specific applications, and provide support for their OoS.