Impact of early estimation of statistical flow features in on-line P2P classification

• Main Authors: Abdalla, B. M. A., Hamdan, Mosab, Khalifa, Entisar H., Elhigazi, Abdallah, Ismail, Ismahani, Marsono, M. N.
• Format: Conference or Workshop Item
• Published: 2020
• Subjects: TK Electrical engineering. Electronics Nuclear engineering
• Online Access: http://eprints.utm.my/id/eprint/92243/; http://dx.doi.org/10.1109/SCOReD50371.2020.9250967294299

Managing high-bandwidth application traffic through identification of bandwidth-heavy Internet traffic is important for network administration. classification based on statistical flow features was proven as an encouraging method for identifying Internet traffic. Early estimation of statistical flow features from first n packets still plays an essential role in accurate and timely traffic classification. In this work, we investigate the impact of early estimation of statistical flow features for on-line P2P classification in terms of accuracy, Kappa statistic and classification time. Simulations were conducted using available traces from the University of Brescia. Results illustrate the early statistical flow features estimation for gives the most significant accuracy and efficiency to detect P2P traffic.