Since 2014 there are more mobile devices on Earth than humans. One of the prominent reasons for the success of mobile devices are mobile applications. On average, more than hundred applications are installed on a single mobile device and millions are available in app stores. To operate properly, 83% of applications available in Google Play rely on the Internet and cause a substantial amount of cellular traffic. Cellular network operators are struggling since years to manage cellular traffic demands caused by the mobile applications – especially at peak times. Users are affected too, because they experience poor network performances.

A popular approach to deal with this problem consists in prefetching mobile application data. This talk will first review prominent state-of-the-art prefetching algorithms and highlight their shortcomings. It will then describe a novel prefetching algorithm that relies on application usage predictions. Lastly, some of the yet-to-overcome challenges will be highlighted.

Paul Baumann is a postdoctoral researcher at the Embedded Systems Lab at TU Dresden, Germany. He completed his PhD in August 2016 with Prof. Silvia Santini as his advisor. Prior to joining the Embedded Systems Lab in 2015, Paul was for three years a research associate at the Wireless Sensor Networks Lab at TU Darmstadt. During this time, he was a visiting scholar at Carnegie Mellon University in Pittsburgh in the group of Prof. Anind Dey. Paul’s research interests lie in the domain of ubiquitous and pervasive computing with a focus on human mobility prediction and mobile sensing.