A Prediction Approach to End-to-End Traffic in Space Information Networks

摘要

Software Defined Space Information Networks (SDSIN) is increasingly applied to life and production, because it can cope with various complex environments and tasks in the future communication environment. In order to better provide users with high-quality services, Internet Service Provider must predict and analyze End-to-End (E2E) traffic. Hence, how to accurately predict the future trend of E2E traffic is great significance to load balance, network optimization and network security. Different from the previous 2-dimension terrestrial network, we study a 3-dimension SDSIN in this paper, then propose an improved Hidden Markov Model (HMM) E2E traffic prediction method. Furthermore, from the practical point of view, we apply the Poisson distribution to initialize the training parameters of the model, and apply some constraints to optimize the predicted values. Finally, the simulation results show that our improved HMM can effectively predict the E2E traffic in SDSIN.