The Cloud and Large Scale Computing (CLSC) group investigates the design and implementation of efficient, scalable, and secure systems and middleware for cloud and decentralized contexts. Our current interests and projects include the topics below. Past topics of interest have also included programmable networks and mobile systems.
We improve the performance, ease of use, and security of decentralized systems. For Ethereum, our recent focus includes integrating networking support for Data Availability Sampling, improving sharding mechanisms and their integration with consensus, and building censorship-resilient rollup designs. We also target the security and dependability of large-scale storage (e.g., IPFS) and service discovery layers. We improve the security and resilience of decentralized protocols, often using Trusted Execution Environments.
Related project: SCEAL, CHIST-ERA project in cooperation with UAIC, Romania and the University of Neuchâtel, Switzerland.
The group develops innovative malware detection systems for diverse targets, including traditional Windows PE malware and emerging targets such as DeFi applications running as smart contracts on blockchains. We also investigate the performance of intrusion detection systems, and the potential benefit of modern hardware in this task.
Related project: Prof. Rivière coordinates the large-scale CyberExcellence project for security-related research in Wallonia and Brussels, funded by the Walloon region.
We design resource allocation, scheduling, and optimization techniques for Big Data systems. Our current focus is on the pair formed by Apache Kafka and Apache Flink. Recent results include efficient resource management for Flink in single and multi-query settings, resource prediction models, and efficient coupling mechanisms based on Envoy proxies.
Related project: GEPICIAD, funded by the Walloon region, in cooperation with Eura Nova.
We study the programmability and performance of distributed cloud systems, i.e., the edge-cloud continuum. We build performance prediction models and schedulers for strict tail-latency objectives. We automate application adaptation and placement, from horizontal cloud setting to hierarchical edge-cloud environments. We study software evolution and understanding of data access models to support this adaptation.
Related project: RAINDROP, ARC project in cooperation with the SPNS group at UCLouvain and the team of Pr. Anthony Clève at UNamur.
We build novel programming and access control models, enabling application developers to build privacy-conscious applications with strictly controlled and user-understandable access to sensitive sensory data. We develop the HubOS operating system for smart homes in this context.
Related project: SSHC (Smart and Social Home Care), funded by Innoviris (terminated)
Finally, the group is active in integrating the research ecosystem in Wallonia and Brussels into European initiatives for shared research infrastructures. It hosts the Belgian node of the Grid5000 network, soon to become part of the larger, European SLICES initiative.
The DIAL institutional library gives access to pre-print of our papers. Only papers published in 2014 and later are listed. You can check the complete list of publications of Pr. Etienne Rivière on DBLP or Google Scholar.