Servicenavigation

Project Details

Back

Projekt Print View

FogStore – A Data Management Platform for Geo-Distributed Fog Environments

Subject Area Security and Dependability, Operating-, Communication- and Distributed Systems
Term from 2018 to 2023
Project identifier Deutsche Forschungsgemeinschaft (DFG) - Project number 415899119
 
Final Report Year 2023

Final Report Abstract

The “FogStore” project aimed to develop a data management system for fog computing, specifically targeting emerging application domains such as the Internet of Things (IoT), autonomous driving, and future mobile networks (5G, 6G). The project focused on addressing the challenges of data distribution, replica selection, and predictive replica placement in geo-distributed fog environments. The project followed a work plan that involved the development of an open-source software prototype called FReD (Fog Replicated Data) as the basis for the project. This prototype implemented key abstractions such as replica nodes, keygroups, and trigger nodes, which proved effective for flexible and efficient data replication in geo-distributed environments. However, the project encountered complexities in researching alternative consensus mechanisms for nodes and had to postpone this aspect to the second year. The project team realized that the trigger node abstraction alone was insufficient to support application developers in creating fog services. As a solution, they developed a lightweight serverless platform called tinyFaaS for managing compute services at the edge. Additionally, they explored predictive replica placement and leveraged machine learning mechanisms, specifically location prediction with Markov models, to anticipate future application requests and optimize replica placement. The project also focused on developing a benchmarking framework for evaluating fog data management systems’ performance in geo-distributed and coordinated execution of parallel workloads. Throughout the project, there were shifts in research attention due to the evolving landscape of mobile networks. Specifically, the anticipation of in-network computing capabilities did not materialize as 5G matured. These were shifted instead to 6G, which also brought along a focus on non-terrestrial low-Earth orbit (LEO) networks. The team extended their research to compute service and data management on LEO satellite networks, leading to fruitful results in terms of publications and the development of a virtual fog testbed tool called Celestial. The project’s main results include the development of FReD, an open-source data management system for the fog with application-controlled replica placement, which provides flexible and efficient replication options based on keygroups and replica nodes. The system offers client-centric data consistency guarantees and supports complex applications through trigger nodes. The project also introduced tinyFaaS, a lightweight serverless platform for the edge that enables elastic scaling of compute services with minimal resource usage. The team explored distributed coordination strategies based on layered coordination systems and investigated predictive replica placement techniques using machine learning models. Additionally, the team made significant contributions to the advancement of data and compute service management in LEO satellite networks with in-network computing capabilities. The project’s findings contribute to the advancement of fog computing research and offer insights into addressing the challenges of data management, replica placement, and coordination in geo-distributed fog environments. The results have potential applications in various domains and lay the groundwork for future investigations in fog computing and mobile network technologies.

Publications

 
 

Additional Information

© 2025 DFG Disclaimer / Copyright / Privacy Policy

Textvergrößerung und Kontrastanpassung