The development of the Internet is dominated by complex interplays between the current Internet technologies and the human user behavior. On the one hand, the ubiquitious Internet influences the daily life, society, as well as industry and business. On the other hand, the continuously growing and changing network requirements call for the development of new Internet technologies to keep pace with the digitizing society. For that, networks have to be flexibly tailored to the needs of the users, and they need to be operated in an effective and sustainable way. At the same time, these technologies have to ensure a high subjective Quality of Experience and satisfaction for the users of the new Internet services and applications.
This is the reason why this research group puts the user to the center of networking technologies. First, technical, individual, and social factors are identified and quantified, which influence the behavior, the perception, and the expectations of users of Internet services. Here, the focus is both on private usage, such as web browsing, video streaming, or augmented/virtual reality, as well as on business applications. Suitable models are derived, which can describe the flexibility of network requirements on different time scales. These analytical, statistical, or machine-learning-based models are the basis to develop intelligent and smart networking technologies, which satisfy the flexible requirements of the users. For this, in particular, the potentials and limitations of machine-learning-supported network management for user-centric communication networks are researched. The performance of the developed technologies is evaluated with the help of formal analysis, simulations, or prototypes, which also ensures their practicability. Here, the business aspects of network operators are quantified as well as the technical, individual, and social gain for users.
- What fundamental concepts drive the formation of Quality of Experience and its interplay with user behavior?
- How to reasonably apply machine learning to network data and network management?
- How to develop a proactive network management for user-centric communication networks?
DFG Emmy Noether Junior Research Group UserNet
(Since October 2022)
In order to allow QoE monitoring for arbitrary Internet applications, the interplays between QoE and user interactions is investigated and modelled based on measurements and subjective studies. In addition, ML methods are adapted to the domain in order to apply them to encrypted network traffic. This allows to quantify the QoE by monitoring interactions and the resulting changes in the encrypted application traffic. Based on this, a data-driven improvement of QoE and QoE fairness is enabled by using reinforcement learning to find optimal network configurations by interacting with the dynamic network environment. By means of powerful, software-defined networking (SDN) technologies like P4, together with available computing resources in the network, such fine-granular models can now be implemented in the network for the first time, such that network management becomes more dynamic. Thus, the implementation of the required ML-based algorithms and components and their integration into network operation is researched.
DFG Group-based Communication
(Since January 2021)
The goal of this research project is analyzing the novel communication paradigm of group-based communication. The findings are used to design edge caching mechanism, which can reduce the network load imposed by user-generated contents, and increase the Quality of Experience of end users.
WINTERMUTE (funded by BMBF)
(April 2020 - March 2023)
This project focuses on AI-based network assessment, policy definition, and enforcement of security in complex networks.
(October 2019 - July 2020)
The objective of the “WebQKAI” project is to infer web QoE key performance indicators (KPIs) from data collected by network devices, which provide insights for operators with respect to network operations and maintenance.
(July 2019 - September 2020)
This project examines the usage of AI methods for the parametrization of convergent, deterministic, industrial networks.
(May 2019 - April 2021)
This project focusses on the relationship between the perceived quality of the performance of business applications by the employees and the technical performance data of such applications.
(since March 2019)
The goal is to monitor and analyze the QoE in the networks of an education service provider, with the eventual goal of improving their services.
(since December 2017 )
What's Up is an interdisciplinary project together with psychologists of the University of Tübingen to analyze the communication of depressive children and adolescents in WhatsApp. With the help of WhatsAnalyzer, an early-warning system for depressive phases will be developed, which can effectively be used in the treatment of depression.
(April 2014 - December 2021)
This project focuses on designing and evaluating new mechanisms in micro tasking platforms to improve the basic concepts with respect to the interests of provider, employer and worker.