EN FR
EN FR


Section: Application Domains

Fog and Cloud computing

Cloud computing has largely evolved in the last years including new networking capabilities as highlighted in the previous section towards the model of XaaS or everything-as-a-service. Moreover, cloud computing continues to be more distributed and aims at integrating more heterogeneous resources. One particular example is fog computing that consists of a massively distributed number of different resources, including low-performance ones. Large network operators have a great interest in fog computing because they already operate such an infrastructure (e.g. a national operator with regional clouds and setup boxes in end users' homes). Softwarization or virtualization of all functions and services will help them to be competitive by reducing their costs. In general, intelligent orchestration of massively distributed resources will be investigated in various application domains, including federated cloud infrastructures, fog computing, 5G networks, IoT and big data infrastructures.

The manageability of such largely distributed systems is a core topic with questions related to monitoring, security and orchestration of resources. Major changes and errors can have dramatic effects on a real system, that actually lead to only minor changes being carried out and slow down innovation and adoption of new propositions. Hence, controlled and reproducible experiments are vital.

As shown by our past work, we are able to quickly adjust to experimental needs in most areas of distributed computing and networking, such as High Performance Computing (HPC), Big Data, Peer-to-peer systems, Grid computing, etc. However, in the context of Resist, we will focus mainly on Software-Defined Infrastructures, gathering cloud computing for compute and storage resources, software-defined networking and network function virtualization for networking. Those infrastructures share many common features: need for performance, for scalability, for resillience, all implemented using flexible software components.

Worth mentioning here is our involvement in the international testbed community (FIRE, GENI). We plan to strengthen our existing links with the Chameleon and CloudLab US projects, to leverage the recently accepted Fed4FIRE+ project on a testbed federation, and, at the national level, to contribute to the SILECS initiative for a new large-scale experimental computer science infrastructure.