부산대학교 도서관

Video streaming has become a prevalent form of entertainment and a vital means of communication, but the challenges of delivering high quality video content over the internet are numerous. One of the key challenges is the varying network conditions that can significantly impact video streaming quality, such as bandwidth fluctuations and packet loss. To overcome these challenges, an adaptive video streaming architecture is needed to adjust the video streaming in real-time to match the changing network conditions and ensure a high Quality of Experience (QoE) for the end-user. This article presents MIGRATE, an orchestrator architecture for video streaming services capable of adapting to user demand in real-time. The study considers an edge/cloud multi-tier network infrastructure. In addition, an Integer Linear Programming (ILP) model and a Greedy solution are proposed to decide the distribution of connections between users and services. Experimental results show that based on the optimization strategy used, it is observed that there is a trade-off between the resources used and the QoE provided to users. Further, we discuss the importance of considering QoE metrics and user engagement in designing video streaming systems.