| Summary: | As encrypted traffic is becoming a standard and traffic obfuscation techniques become more accessible and common, companies are struggling to enforce their network usage policies and ensure optimal operational network performance. Users are more technologically knowledgeable, being able to circumvent web content filtering tools with the usage of protected tunnels such as VPNs. Consequently, techniques such as DPI, which already were considered outdated due to their impracticality, become even more ineffective. Furthermore, the continuous regulations being established by governments and international unions regarding citizen privacy rights makes network monitoring increasingly challenging. This work presents a scalable and easily deployable network-based framework for application identification in a corporate environment, focusing on video applications. This framework should be effective regardless of the environment and network setup, with the objective of being a useful tool in the network monitoring process. The proposed framework offers a compromise between allowing network supervision and assuring workers’ privacy. The results evaluation indicates that we can identify web services that are running over a protected channel with an accuracy of 95%, using low-level packet information that does not jeopardize sensitive worker data.
|
|---|