According to IDC studies “Digital universe” published in December 2012, the world has created or reproduced around 2.8 zettabytes (1021 bytes) of data, and we forecast around 40 zettabytes over a five-year horizon (equivalent to 47,200 billions of billions of bytes).
Those data are the raw material of IT: we must stock, classify and index them in order to retrieve, manipulate and analyze their content to extract the meaning, while processing performance issues induced by their large volumes.
In the 2000s, the sharp fall of disks prices made data storage as a secondary challenge. Currently, the fast growth of large data volumes has made storage management once again a major concern for large datacenters.
Managing large volumes of data requires finding a compromise between various parameters:
- Capacity storage to preserve the information mass
- Velocity to adapt the I/O speed of media to the applications
- Electricity consumption, since energy supply is a major constraint on large sites
- Media reliability
- Costs, etc.