Why a cloud?
The choice to use cloud technology was made in light of the experience gained from some 12 years of practice using distributed infrastructures for high-performance computing in the domain of the life sciences, ranging from clusters to grids and finally terminating with the cloud in 2007. Since 2008, the IFB-GRISBI, “Groupe de Réflexion sur les InfraStructures Bioinformatiques” (Think Tank on Bioinformatics IT Infrastructures), evaluates this technology and establishes the guidelines.
A cloud is based on computer resource abstraction technologies (CPU, storage, RAM, network). Different implementations can be considered, ranging from the supply of material resources to the implementation of software containers on a standard operating system and including the creation of virtualization layers and the use of virtual machines. According to the American National Institute of Standards and Technology, clouds should exhibit the following properties:
- Resource pooling
- On-demand self-service
- Broad network access
- Dynamic resource scheduling (elasticity)
- Measured service