Den 27 november var det dags för doktoranden Stefan Forsström att försvara sin avhandling för betygsnämnden. Stefan har genom sin forskning varit med och tagit fram plattformen Sensible Things som gör det möjligt att genom Internet of Things koppla upp emot 50 miljoner enheter samtidigt och snabbt kommunicera olika mätvärden. Som ett erkännande för sin forskning, har Stefan fått pris för både Best Paper på en konferens och utmärkelse för bästa studentuppsats.
Date: Thursday 27th of November 2014
Time: 10:15 – 13:00
Place: Room L111, Campus Sundsvall
Opponent: Prof. Antonio Corradi, University of Bologna, Italy
Prof. Shaogang Gong, Linköping University, Linköping, Sweden;
Dr Claus Popp Larsen, Swedish ICT Acreo, Stockholm, Sweden;
Assoc. Prof. Aron Larson, Mid Sweden University, Sundsvall, Sweden
The focus of this thesis is the enabling of Internet-of-Things (IoT) applications that exchange sensor and actuator information on a global and ubiquitous scale. The main challenge is to create a common architecture suitable for different types of IoT applications. So far, proprietary and centralized solutions have dominated related work, although an open and distributed approach offers better support for a global IoT. Therefore, the focus in this thesis is on fully distributed middleware approaches and architectures for disseminating information on a global scale between ubiquitously connected entities. In addition to this, there is an impending information overload that needs to be avoided, where the amount of produced sensor information greatly succeeds the usefulness and possible utilization of each individual value. The problem is split into three concrete goals, where the first one surveys the current trends for creating IoT applications. In order to determine their demands on the underlying support and propose an approach for enabling next generation services. The second goal is to implement the proposed approach as a fully functional application platform. Finally, the last goal is to suggest mechanisms to manage resource limitations in order to avoid an information overload. The obtained results include an implemented IoT platform with ubiquitously connected devices that can globally share information between each other in a scalable manner. The platform utilizes a fully distributed architecture and is extended with low cost protocols that only communicate relevant information when it is needed, thus improving the longevity and performance of services based on the IoT. Therefore, the main contribution made by this thesis is that fully distributed architectures outperform other architectures when it comes to global and ubiquitous IoT services. The impact of this work is thus that it highlights the problems associated with poorly scaling architectures, which might be profitable as walled garden solutions but are inferior when it comes to global IoT services.
About the respondent
Stefan Forsström was born on the 11th of May 1984 in Sundsvall, Sweden. He achieved his Master of Science and Engineering degree at Mid Sweden University in August 2009, which also won the Sundsvall 42 scholarship for excellent thesis work. Directly after, Forsström started as a PhD student at Mid Sweden University where he focused on sensor based context-aware applications. He presented his licentiate thesis in April 2011, which focused on enabling these sensor based services on mobile devices using context views. Proceeding this, Forsström’s research work has been focused on enabling ubiquitous sensor-assisted applications on lightweight devices, for a global Internet-of-Things. As a recognition for his research, Forsström has been awarded both a best paper award and a best student paper award.