In a context where data production and usage is mututating at a fast pace, the volume of data is increasing at a hardly conceivable speed. In a few years, a software application can mutiply the data it produces by orders of magnitude. Scientific applications are a good example, such as in the domain of astronomy with the increase in the number and capacity of sensors1. In every days life, digital environments are fast-growing (Web, social networks, e-commerce, e-administration, open data, services, smart buildings and smart cities, Web of Things, ...). Networks connect proximity devices (sensors, plugs, smart-phones, tablets, personal computers, connected devives, robots, ...) to infrastrutures, more and more powerful and distant (servers, clusters, grid, cloud). All these elements embed, use, produce and transmit data in various objectives including observing and memorizing uses and behaviors (user traces in services, measurement in cars, planes, ...). Stakes related to these data cover all levels of society: individual, socio-economic, strategic, etc. Having data is necessary but insufficient to meet the corresponding needs. Their utilization requires to integrate them, to query them in reasonnable time, to compare, aggregate and analyse them in order to extract information and knowledge for them, for example in order to assess their quality.
Our scientific objectives are about management of large volumes of data, but also heterogeneous data related to non conventionnal sources (streams, services and activators). While easing acces, interrogation, analyse and exploration of data, solutions have to guarantee good properties from the points of view of data and answer quality, security and privacy. Indeed, if having new services related to data exploitation is interesting, centralization and monopolization of data must not be the only, imposed, solution. Quite the opposite, it is crucial to improve the control of data (for individual and for organizations) as well as their access, diffusion and usages.
The group is organized around the following three thematics:
Applications, such as smart building, allow members of the group to get together across those three thematics.
The LSST project envisage to produce 14 Tb of data per night. ↩︎
