IBM new 120PB Mega data repository

On August 29, according to foreign media reports, an article in the MIT technical review last week pointed out that IBM researchers are working on a new 120PB (1015) byte data repository to meet 200,000 regular hard drives at the same time.　　The Mega data memory will store 1 trillion of files and provide sufficient computational space for complex system simulations, such as weather model calculations. The new system uses IBM's common parallel File system (GPFS), which enables supercomputers to access data more quickly, transfer a single file across multiple disks, and read and write multiple pieces of information at the same time.　　GPFS leverages the cluster architecture to provide faster file data access, automates transmission across multiple storage devices, and optimizes performance with existing storage space, as well as IBM Watson's storage engine. Its 120PB storage space is equivalent to 24 billion 5 trillion MP3 files. Take Facebook's 2009-year data as an example, storing data for 25TB a day, so it can only store 4,915 days of data. With network data and other offline data now growing in series, 120PB may not be able to meet existing needs in 5 years, and still need further innovation.