Decentralized Systems and Network Services Research Group
Under the term "Decentralized Systems and Network Services" we understand distributed and networked technical systems that span over more than one administrative domain. Hence, their operation depends on more than one party. Our research focuses on:
- Blockchains, consensus and peer-to-peer networks
- Network Security Monitoring
- Secure and privacy-aware computing in partially trustworthy environments.
News | News Archive
In the upcoming semester, we offer a new course: The Research Focus Class, that combines a seminar with a lecture. In the summer term 2021 we will address aspects of the highly relevant and up-to-date area of Blockchain and Payment Channel Networks. In the lecture as the first part of the course, we will teach basics in this area and in the second part of the course, the seminar, students will work on advanced topics. Register here.
In the Open Access journal IEEE Access, our article Analysis of the Matrix Event Graph Replicated Data Type was published. With the aim of application in decentralized publish-subscribe systems, in the paper, we extract and analyze the Matrix Event Graph(MEG) from the Matrix middleware, which is primarily used there to depict the causal history of chat messages. We show that the MEG is a so-called Conflict-Free Replicated Data Type, and thereby provides Strong Eventual Consistency. In addition, we show that the MEG, via its reduced notion of consistency when compared to traditional Distributed Ledger Technologies, can cope with environments where the majority of the participants exhibit byzantine faults. We close the article with an analysis of the scalability in form of the number of parallel causal chains in the MEG, for which we can expect an almost optimal evolution according to our results.
We present a demonstration titled Practical Trade-Offs in Integrity Protection for Binaries via Ethereum at this year's ACM/IFIP Middleware (Video). We show the practicability of our concept and how revocations of software binaries can be communicated to users in a timely manner. We also briefly touch on trade-offs that arise in a practical deployment.
At Virtual Zeek Week 2020, we presented the Emojifier and the current state of our work on the Packet Analyzer Framework for Zeek. The framework enables the flexible addition of protocols at lower network layers and thus opens up new possibilities for monitoring networked systems, for example in the context of critical infrastructures.