Section: Application Domains

Research domain

To address the evolution of distributed platforms in recent years, we focus on the following areas:

• Distributed algorithms for dynamic and large networks. Network topology is no more static; distributed systems are increasingly dynamic, i.e., nodes can join, fail, recover, disconnect and reconnect, and change location. Examples include IaaS cloud computing infrastructures, where virtual machines can be moved according to load peaks, opportunistic networks such as DTNs (Delay-Tolerant Networks), and networks of robots.

• Management of distributed data. In emerging architectures such as distributed hash tables (DHTs) and cloud computing, our research topics include replica placement, responsiveness, load balancing, consistency maintenance, consensus algorithms, and synchronisation. This research direction is funded by several new collaborative projects (ConcoRDanT, MyCloud, Nu@age, Odisea, Prose, Shaman, Spades, Streams, R-Discover) and by industrial funding (Google).

• Performance and robustness of Systems Software in multicore architectures. Our research focuses on the efficient management of system resources at the user level. Issues considered include efficient synchronization and memory management in large-scale multicore architectures. At the same time, we focus on the robustness of systems software, based on the Coccinelle technology. This work is funded by ANR ABL and InfraJVM.