IFB academic Cloud
The IFB cloud is designed to take into account the specificities of the life science domain including the use of a large number of data collections (more than a thousand exist) with their format heterogeneity and broad geographic range. In addition, the analysis of biological data requires the use of several hundred bioinformatics tools with different types of interfaces (graphic user interfaces, web portals, web services, and command line interfaces). IFB’s objective is to provide scientists and engineers as well as users and developers with the required computing and storage capacity while providing a flexible, user-friendly, easily adjustable solution.
Currently in a pilot stage, the IFB cloud comprises 200 computing cores, two terabytes (TB) of RAM and 50 TB of storage. We plan to increase these resources to 3,000 cores with 500 TB of storage by the end of 2016 and, finally, achieve 10,000 cores with two petabytes (PB) of storage in 2017. The IFB’s IT infrastructure is hosted at IDRIS, one of the major national high-performance computing centers (GENCI). This ensures the stability of the IFB’s infrastructure by providing a reliable power supply, cooling system and large network bandwidth commonly found in such environments.
The long-term objective is to create a federation of clouds that rely on the interconnected IT infrastructures of the IFB’s platforms, providing personalized services to analyze life science data.The organization model considered will create a balance between data movement, which may be very costly in some cases and even impossible with confidential data, and the automatic deployment of appliances on local clouds, the closest possible to data repositories but requiring a compatible local infrastructure.