Mobile networks underpin numerous vital operations of the modern society and are arguably becoming the most important piece of the modern communications infrastructure. The use of mobile networks has exploded over the last few years due to the immense popularity of mobile devices such as smartphones and tablets, combined with the availability of high-capacity Wifi and 4G/LTE networks. While today’s mobile networks do a reasonably good job connecting people, they face a number of challenges when it comes to supporting the diverse demands of emerging applications, simply because they were not designed with their requirements in mind. The applications and services of the future have quite different requirements in terms of capacity, latency, reliability, and efficiency. For example, while self-driving cars require very low delay with a very high reliability, for future media applications (e.g., ultra HD, 3D video) very high capacity is required. Therefore, the vision for future mobile (e.g. 5G networks) networks is not only providing higher capacity and an improved perceived user experience for different applications and services but also enabling application designers and network architects to build end-to-end virtual networks tailored to their applications’ requirements. The promise of great flexibility to support many different services and applications, however, comes at the cost of increased complexity. Real-world network measurements are essential in many network investigations, especially for performance and reliability analysis of complex networks and, applications and services that are running on these networks, since they provide an environment that is hard to mimic in models and simulators.

The main goal of Mobile Systems and Analytics (MOSAIC) is to empirically analyze mobile systems in order to design and validate novel protocols and applications for future mobile systems. To achieve its goal,  MOSAIC follows a “hands-on” approach using open platforms such as MONROE to identify the bottlenecks in the current mobile networks (e.g. 3G/4G/5G, Wifi and satellite) and leverage on this knowledge to design and validate novel approaches for future mobile networks. To improve our understanding of mobile systems and the performance of applications running on them, MOSAIC focuses on merging the domain knowledge with machine learning methodologies. Building on the application-specific measurements and metrics, MOSAIC designs and validates application-specific protocol and system enhancements for mobile systems. More specifically, MOSAIC leverages on the experimental evaluation and characterization of mobile networks to design: (i) novel protocol extensions (e.g. TCP/MPTCP, QUIC/MPQUIC) and (ii) robust multimedia (e.g. DASH and HEVC) applications over mobile networks, especially considering challenging mobility scenarios such as the ones experienced on trains and drones.