JORDAN: A Novel Traffic Engineering Algorithm for Dynamic Adaptive Streaming over HTTP

Video streaming accounts for most of the Internet traffic. To cope with the increasing traffic and achieve better service, Internet service providers (ISPs) and applications take actions to mitigate the problem independently. Namely, ISPs perform traffic engineering (TE) to ensure the whole network capacity is evenly utilized based on an estimated traffic matrix. Additionally, many applications employ rate adaption mechanisms (such as Dynamic Adaptive Streaming over HTTP, DASH) to adjust streaming bit rates according to the measured network conditions. However, deploying these two interrelated methods separately may lead to suboptimal performance. To solve this problem, we first illustrate the challenges. Then we formulate a problem for Jointly Optimizing Routing and Dynamic Adaptation in Network (JORDAN) and propose an algorithm to obtain a near-optimal solution. Finally, our simulation results show that JORDAN can achieve up to 180% improvement in terms of average bit rates than TE and DASH working separately.