Decentralised Knowledge

Decentralized edge systems represent a transformative shift in computing paradigms by redistributing processing capabilities closer to the data generation points. This architecture ensures low-latency responsiveness, crucial for applications demanding real-time interactions such as autonomous vehicles and augmented reality. Furthermore, by processing sensitive data locally, these systems bolster privacy and security, addressing growing concerns in our interconnected world. The scalability of decentralized edge systems is evident as they efficiently handle increasing device connections, making them ideal for the expansive realm of the Internet of Things (IoT). Additionally, the dynamic workload distribution and redundancy features enhance reliability, contributing to a more resilient computing infrastructure. Embracing decentralized edge systems is not just a technological evolution but a strategic response to the demands of a rapidly advancing digital landscape.

The aim of this work will be to utilise the aforementioned advantages by means of simulations, emulations and, if necessary, real applications using efficient epidemic algorithms and protocols. One of these algorithms has already been initially published by my esteemed colleague Sashikant Ilager under the name DEMon [1]. DEMon will be matured in the course of this work. In addition, further programs are being researched that not only enable monitoring and management in decentralised edge environments, but can also further acquire, distribute and analyse the overall knowledge of the system and also the more specific knowledge of an individual node in a decentralised manner. DEMon provides initial results on how such algorithms can operate robustly in edge systems and how such a prototype, including a test environment, can be implemented. Now the initial questions are: 

• How can DEMon be extended to become more efficient/robust?
• What changes or additions are necessary to use a similar approach to efficiently exchange more and different knowledge?
• How should insights be recognised and also processed or stored?
• What alternatives are there to exchange information efficiently and robustly in volatile edge systems?

DEMon [1]

Edge computing supports IoT at the network edge, requiring efficient monitoring. Traditional centralized systems create latency and reliability issues in volatile edge environments. To address this, we propose DEMon, a decentralized monitoring system using a stochastic gossip protocol. DEMon ensures fast, reliable data access, adaptively managing monitoring parameters. Implemented as a lightweight, portable container system, DEMon efficiently disseminates and retrieves monitoring information, addressing edge monitoring challenges.

[1] S. Ilager, J. Fahringer, S. C. d. L. Dias and I. Brandic, "DEMon: Decentralized Monitoring for Highly Volatile Edge Environments," 2022 IEEE/ACM 15th International Conference on Utility and Cloud Computing (UCC), Vancouver, WA, USA, 2022, pp. 145-150, doi: 10.1109/UCC56403.2022.00026.