Enriching Network Traffic Traces for Increasing Classification Accuracy

Обогащение трасс сетевого трафика для повышения точности классификации
Sripath Mishra, Akhil Prasad, Sonia Fahmy, Sonia Fahmy
2025-11-15

FlowPicapplication fingerprintingnetwork traffic classificationpacket size sequencestraffic trace augmentation
Internet service providers often fingerprint the type of application, e.g., video streaming or video conferencing, from a network traffic trace, for security and traffic engineering purposes. Several traffic classification methods have been proposed to apply machine learning techniques to packet timestamp and packet size information in traffic traces to perform such fingerprinting. However, classification accuracy drops when the machine learning methods are trained on traces that do not capture diverse end point and underlying network conditions. In this work, we compare a number of augmentation techniques that attempt to mitigate this problem, and evaluate their effectiveness when we systematically change specific end point and underlying network conditions. We find that the Rosetta traffic classification method that uses packet size sequences benefits much more from current augmentation methods than the mini-FlowPic method that uses both packet timestamps and sizes. We believe that there is a critical need for customized augmentations that enrich traffic traces, while being faithful to behaviors of real users and real protocols.1
1
Machine-learning traffic classification accuracy declines when training traces lack diverse endpoint and underlying network conditions.
2
Rosetta, which uses packet-size sequences, benefits substantially more from existing augmentation methods than mini-FlowPic, which uses packet timestamps and sizes.
3
The findings indicate a need for customized trace augmentations that enrich data while preserving realistic user and protocol behavior.
4
The study compares multiple traffic-trace augmentation techniques under systematic changes to endpoint and network conditions.

Internet application network traffic traces under diverse endpoint and underlying network conditions

the effectiveness of traffic-trace augmentation techniques for improving application classification accuracy and robustness to changing conditions

Publication Details
Publication Date
2025-11-15
Journal
Publisher
ISSN
Cited by
0
Access Type
Author Information
Authors
Sripath Mishra
Akhil Prasad
Sonia Fahmy
Sonia Fahmy
Explore further
Open the scid.ai AI chat with a ready-made request: it will find papers on a similar topic and help build a literature review.
Find similar papers in the chat →
Make a presentation
100%