Research on Application Identification and Feature Extraction Algorithms Based on DPI
Tao Liu, Zhihong Feng, Ruohua Jin · 2025
Network application identification is crucial for traffic management and user behavior analysis. While peer-topeer (P2P) applications offer convenience, they also contribute to network congestion. Therefore, effective P2P traffic identification and management are pressing concerns. This paper proposes an offline detection algorithm, based on the Apriori algorithm, that addresses the challenge of extracting non-continuous features. The algorithm's effectiveness is validated through probabilistic analysis. For application identification, hash mapping based on Apriori-extracted features enables rapid identification, with correctness also verified using probabilistic analysis.