NSF Workshop on Ultra-Low Latency Wireless Networks

Arizona to address challenges in the design of ultra-low-latency wireless networks. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation. Wireless networks have become a ubiquitous part of everyday life all around the world. Yet, wireless networks are very unpredictable on one critical aspect: communication delay. It is well-known and widely observed that the delays incurred in accessing a wireless network can vary widely. The issue of unpredictable and often high latencies precludes wireless networks from being used in mission-critical environments. Even common applications, such as real-time video-conferencing, suffer from poor performance due to the unpredictable nature of these delays. Today's communication networks are largely geared towards latency tolerant (web, chat, email) content. Thus, these networks have been typically engineered with a focus on improving network capacity, with little attention to delay performance. However, in a range of domains, a new wave of socially useful applications are emerging based on automated sensors and actuators operating in closed-loop or open-loop control systems. In these systems, including internet of things (IoT) applications, vehicular networks, smart grid, distributed robotics, and other cyber-physical systems, the requirements for latency could be two or three orders of magnitude more stringent than traditional applications. In addition, there are immersive services like online gaming and augmented reality that also require latency much smaller than what is achievable in today's wireless systems. Furthermore, some of these emerging applications often require deployment at a larger scale, making it even more challenging to provide ultra-low latency over the network. Section 4 of the report discusses emerging applications and their latency requirements. In order to facilitate the emergence of such " mission critical " low-latency application, it is imperative that the research community develops new mechanisms for enabling consistently low-latency wireless networks. Technical challenges in achieving end-to-end low-latency exist at various levels. Section 2 of the report outlines both challenges as well as opportunities that need to be addressed in various domains, including network control algorithms for reducing latency (e.g., scheduling, routing), and Information Theoretic methods for improving delays at the physical and MAC layers. At the chip level, bottlenecks for designing ultra low-latency wireless NoCs are bandwidth and resource limitations of the physical layer (e.g., device technologies, wireless transceivers, on-chip routers, and antennas). Finally, much of the delay in today's networks arises …