Team leader: Ga√ęl THOMAS

The Benagil team studies how to design more efficient and safer distributed systems with a focus on their core components (hypervisors, operating systems, language runtimes, data storage and communication libraries). The team targets systems used in large-scale computa- tion for which efficiency and safety is paramount, such as large-scale distributed data stores, machine learning frameworks, data analytics platforms, and high-performance applications. To improve the performance and safety of these large-scale distributed systems, Benagil devises performance profiling and analysis techniques that scale both with the size of the system and the amount of collected data. Based upon this analysis, Benagil improves both the efficiency and safety of their components in the cloud. The team also proposes new system mecha- nisms and interfaces for two emerging programming models: serverless computing and edge computing.