TPT: A Scalable Traffic Path Tracking Scheme Using Improved Viterbi Algorithm in Satellite Internet

Satellite Internet (SI) is an essential component of 6G as an effective supplement to terrestrial Internet in the future. Owing to its relatively limited processing power and bandwidth, distributed denial of service (DDoS) attacks may cause substantial damage. Moreover, the complexity and difficulty of defense also increase due to the constantly changing topology scale of SI when supporting various applications. Therefore, path tracking must be maintained with high precision at different scales to realize protection at the boundary. However, the marking rate limits traditional packet marking methods, causing a decrease in the tracking accuracy when the topology is expanded. Therefore, we propose a scalable traffic path tracking (TPT) scheme. Firstly, a lightweight description of malicious traffic in the entire network is realized by improving a multi-dimensional traffic feature digesting method. Secondly, the connection relationship sparse matrix of nodes is obtained based on the Hidden Markov Model modeling of the network topology and attack scenarios. These generate the observation probability matrix B and transition probability matrix A of the traditional Viterbi algorithm. Finally, we optimize the Viterbi algorithm by adding an index matrix of the next-largest probability value to eliminate tracking loops, thereby achieving high tracking accuracy for topologies of different scales. The Keysight Ixia platform is used to generate malicious traffic in the experiments. The results demonstrate that the scheme can maintain a tracking accuracy of over 99% against various DDoS attacks in topologies of different scales, which is more accurate and scalable than existing methods.