A video-spam detection approach for unprotected multimedia flows based on active networks

This work presents a real-time protection mechanism for MPEG-4 flows against video spam. Applications like Internet telephony, video on demand or life video streaming become more and more important. Because of the increased appearance of e-mail spam during the last years, it is expected that these services will get into the focus of the spam industry. Since it is necessary to cope with the individual QoS-requirements of media flows and the characteristics of varying transmission media, it might be advisable not to encrypt media flows such that intermediate and edge nodes could optimize the transmission of these flows, e.g., by protocol boosters. Moreover, currently used video-conferencing systems and real-time multimedia applications mostly lack integrated security features. We have been able to demonstrate the blending of media streams with unwanted content and have implemented an active networking service that analyzes MPEG-4 media flows. Additionally, it is capable of detecting and blocking the above mentioned attack.

1. Introduction

Applications like Internet telephony, video on demand or life video streaming gain more and more in importance. For example, a video-conference system allows to discuss in real-time with colleagues of a remote branch, without having to travel. Though interactive applications pose strict requirements with regard to end-to-end latency (maximally 400 ms) and jitter, it is often not possible to integrate security services as data authentication, integrity and confidentiality.