NGNI, Headerbild, FUSECO FORUM 2016
Thu., Nov. 03, 2016 to Fri., Nov. 04, 2016 , Fraunhofer-Forum, Berlin

Sergio Barbarossa

Sapienza Univ. of Rome, Italy

Sergio Barbarossa is a full professor in the Information Engineering, Electronics and Telecommunications Department of Sapienza University of Rome, Italy, where he leads the Signal Processing for Communications Lab. 
He is IEEE Fellow and EURASIP Fellow. He has been IEEE Distinguished Lecturer and received the EURASIP Technical Achievements Award in 2010. He was the recipient of the IEEE Signal Processing Society Best Paper Award for the years 2000 and 2014. For more than 25 years Prof. Barbarossa has been working in signal processing, radar remote sensing, sensor and telecommunication networks. Since 2000 he has been deeply involved in European projects. His main current research interests are on mobile edge computing, 5G networks, distributed optimization, machine learning and signal processing on graphs. 

Abstract

Enabling low latency mobile edge computing through millimeter-wave links

Mobile Edge Computing (MEC) brings information technologies closer to mobile users by empowering radio base stations with additional computational capabilities. Even if the single base station may have limited computing or caching capabilities, the coordination of multiple base stations can give rise to a powerful and scalable distributed system able to deliver services with low latency. 

A further major thrust to this framework comes from the usage of millimeter-wave (mmW) communications as a wireless backhaul between base stations or even for the mobile access. In this talk, we will show how MEC can benefit from mmW in optimizing computation offloading from the mobile user to nearby base stations, minimizing the energy necessary to run mobile applications remotely under strict latency constraints. We will also explore how these benefits vary as a function of the carrier frequency, incorporating imperfect knowledge of the steering angles used for beamforming and intermittent channel models, suitable for mmW beam forming channels.