Host: Elizabeth Belding
Title: Data-Driven End Intelligence in the Mobile Networked Systems
The 4G/5G mobile networked systems are providing "anywhere, anytime" Internet access for billions of users. In the foreseen future, they are also envisioned to serve new applications (virtual/augmented reality, self-driving cars, drones, and more). However, the current systems still rely on the "black-box" operations for the network infrastructure. Consequently, the end devices lack intelligence on the underlying network.
In this talk, I will present a data-driven approach to enhancing end intelligence in the 4G/5G systems. I will show how software-hardware cooperation enables fine-grained network data that were not possible in existing solutions. I will next demonstrate how to leverage these data to construct analytics in order to catalyze end intelligence. Last, I will describe how the enhanced end intelligence, in return, helps the 4G/5G infrastructure with more provable reliability and efficiency. These results make a case for the "Knowledge Plane" for the next-generation mobile Internet.
Yuanjie Li received his Ph.D. in Computer Science from UCLA in 2017. He is the co-founder of MobIQ Technologies, a startup that offers vertical mobile network analytics for mobile operators and application vendors. His research interests include networked systems, mobile computing, and network security. He is a recipient of ACM MobiCom’17 and MobiCom'16 Best Community Paper Awards, UCLA Dissertation Year Fellowship in 2016, and Outstanding Chinese Oversea Graduate Student Award in 2016. His work has resulted in an open-source community tool (MobileInsight), which has been used by 268 universities and companies since its release in 2016.