Abstract:
With the ever increasing growth in the number of online video viewers on internet, the service providers are getting curious regarding the nature of content being revolved over the network without actively interacting with the network device in order to take action rapidly in support of their diverse security business objectives that may include detecting the pornographic videos, video cyber bullying, offensive as well as fake videos, most viewed content, user’s internet profile and video content similarity etc. Due to the dispersion of encoded video streaming techniques, the network video traffic classification has turn out to be a challenging task to perform as devoid of the authentic decryption key, it is obstinate to comprehend the actual content viewed by the user. However, the current advances in machine learning has demonstrated the fact that encryption can also lead to certain information leak which yields promising results in determining the actual transmitted content. This research exploits the classical machine learning algorithms to propose a classifier truly for determining the encrypted video content sighted by users over diverse video sites for instance YouTube, Netflix and Dailymotion under the normal network conditions. This classifier foretells the content with an accurateness of more than 98% within a second and has the ability to execute all the network administration and sanctuary related business objectives.
