Traffic pattern-based content leakage detection for trusted content delivery networks

Hiroki Nishiyama, Desmond Fomo, Zubair Md Fadlullah, Nei Kato

Research output: Contribution to journalArticlepeer-review

6 Citations (Scopus)

Abstract

Due to the increasing popularity of multimedia streaming applications and services in recent years, the issue of trusted video delivery to prevent undesirable content-leakage has, indeed, become critical. While preserving user privacy, conventional systems have addressed this issue by proposing methods based on the observation of streamed traffic throughout the network. These conventional systems maintain a high detection accuracy while coping with some of the traffic variation in the network (e.g., network delay and packet loss), however, their detection performance substantially degrades owing to the significant variation of video lengths. In this paper, we focus on overcoming this issue by proposing a novel content-leakage detection scheme that is robust to the variation of the video length. By comparing videos of different lengths, we determine a relation between the length of videos to be compared and the similarity between the compared videos. Therefore, we enhance the detection performance of the proposed scheme even in an environment subjected to variation in length of video. Through a testbed experiment, the effectiveness of our proposed scheme is evaluated in terms of variation of video length, delay variation, and packet loss.

Original languageEnglish
Article number6463398
Pages (from-to)301-309
Number of pages9
JournalIEEE Transactions on Parallel and Distributed Systems
Volume25
Issue number2
DOIs
Publication statusPublished - 2014 Feb 1

Keywords

  • Streaming content
  • degree of similarity
  • leakage detection
  • traffic pattern

ASJC Scopus subject areas

  • Signal Processing
  • Hardware and Architecture
  • Computational Theory and Mathematics

Fingerprint

Dive into the research topics of 'Traffic pattern-based content leakage detection for trusted content delivery networks'. Together they form a unique fingerprint.

Cite this