IBM Develops World's Largest Data Warehouse

Every year we hear that one of the corporations is building another supercomputer by order of the Pentagon, or some research company. The other day there was news of this type, only now it is not about creating a supercomputer, but about creating the world's largest data storage system. The development of this system, consisting of approximately 200,000 hard drives, is being carried out by IBM. The amount of data that can be stored in such a system is 120 petabytes.

This storage system is created not just like that, but by request. Unfortunately, the customer is not known yet, and it would be interesting to know who and why needed such a data storage, right? Now it is only known that the repository is likely to be used to store data related to climate models, as well as data from various medical and biological studies. Perhaps, the system is being created by the order of several companies having common interests.
')
It is also argued that for this data warehouse, experts developed their own GPFS-based file system, which allows you to store parts of the same file on multiple media. GPFS allows you to achieve a read / write speed of several terabytes per second. More information about this system yet.

It is also known that the cooling in this system is liquid, and each hard disk will be cooled separately. So far, there are several difficulties that developers still have to overcome. One of these problems is the creation of a reliable technology that allows the entire system to work flawlessly, even when several hard drives fail.

Via engadget