Analysis of YouTube DASH Traffic

To design efficient resource allocation algorithms that satisfy users' quality of service requirements, it is necessary to know or correctly estimate the traffic characteristics. In recent years, the share of video streams in the overall traffic has reached 70% and continues to grow. Most of this type of traffic is generated by popular video hosting services, such as YouTube, NetFlix, and others, which use the Dynamic Adaptive Streaming over HTTP technology. This paper discusses the results of a study on the properties the most popular videos on YouTube, the world's largest video hosting service. In addition, it is shown how traffic parameters depend on the application used to watched Youtube videos and operating systems of PCs and mobile devices.