- a lecture on Latency guarantees and service quality in 5G within the 5G System classes (master in ICT) from 10:30 to 12:00 in room 318 DEI/G
- a research talk on Classifying flows and buffer state for YouTube's HTTP adaptive streaming service in mobile networks at 14:00 in room 201 DEI/A.
Latency guarantees and service quality in 5G
Latency guarantees are essential for services such as video-telephony, connected cars and industrial control. While previous mobile networks were designed for voice and best-effort data, 5G will be the first generation offering millisecond-latency to generic service classes. This lecture will walk through the link-layer mechanisms behind 5G’s Ultra-Reliable Low Latency Communication (URLLC) mode. We will focus on scheduling disciplines, mechanisms for queue control, and the necessary changes in frame design that will be behind 5G’s network slicing. Besides exploring specific designs for 4G and 5G, we will discuss fundamental problems and tradeoffs, comment on performance measurement, and provide an update on the state of URLLC in 5G standardization.
Classifying flows and buffer state for YouTube's HTTP adaptive streaming service in mobile networks
Accurate cross-layer information is very useful to optimize mobile networks for specific applications. However, providing application-layer information to lower protocol layers has become very difficult due to the wide adoption of end-to-end encryption and due to the absence of cross-layer signaling standards. As an alternative, this paper presents a traffic profiling solution to passively estimate parameters of HTTP Adaptive Streaming (HAS) applications at the lower layers. By observing IP packet arrivals, our machine learning system identifies video flows and detects the state of an HAS client's play-back buffer in real time. Our experiments with YouTube's mobile client show that Random Forests achieve very high accuracy even with a strong variation of link quality. Since this high performance is achieved at IP level with a small, generic feature set, our approach requires no Deep Packet Inspection (DPI), comes at low complexity, and does not interfere with end-to-end encryption. Traffic profiling is, thus, a powerful new tool for monitoring and managing even encrypted HAS traffic in mobile networks.
Stefan Valentin is Professor for Mobile Networks and Foundations of Computer Science at the Darmstadt University of Applied Sciences since Oct. 2018. Stefan graduated in EE with excellence from the Technical University of Berlin, Germany in 2004 and received his Ph.D. in CS with summa cum laude from the University of Paderborn, Germany in 2010. In the same year, he joined Bell Labs, Stuttgart, Germany as a Member of Technical Staff, where he worked on wireless resource allocation algorithms for 4G and 5G. From 2015 to Sep. 2018 he was with Huawei's Mathematical and Algorithmic Sciences Lab in Paris as Principal Researcher and team leader. Stefan’s research interest is wireless resource allocation and load balancing for mobile networks, based on methods from mathematical optimization and machine learning. Stefan’s algorithms are deployed in base stations around the world and more than 30 patent applications have been filed for them. Besides several awards from industry, he received 2 best paper awards, the Klaus Tschira Award in 2011, and IEEE ComSoc's Fred W. Ellersick Prize in 2015.