In 2011, the LIGO Laboratory needed to migrate 2.9 petabytes of data representing 12 years of observations to an up-to-date, reliable and scalable storage system that could automate access to existing data archives as well as support significantly higher data rates of up to 1 petabyte per year from the next generation Advanced LIGO experiment.
To support rapid data growth, perform deep data archiving, and simplify their IT environment, the LIGO Lab upgraded from Oracle’s StorageTek T9940B tape drives at two LIGO observatories and from Oracle’s StorageTek T10000B tape drives at the central Caltech repository to Oracle’s StorageTek T10000C drives throughout. This delivered 25x the capacity and up to 8x the performance at the observatories and 5x the capacity and up to 2x the performance at the central repository.
With 25 percent greater capacity on a single cartridge than any other tape drive, the StorageTek T10000C tape drives and StorageTek SL8500 and SL3000 modular library systems help Caltech consolidate hardware in its data center to make room for increased capacity requirements of up to 1 PB per year, and store millions of large files, in addition to more than 500 million small files.
Using Oracle’s StorageTek tape drives in this project allowed the LIGO Lab to double its data throughput up to 252 megabytes per second, providing the necessary performance to manage, analyze and support vast amounts of new and historical data.
The LIGO Laboratory is also using 400 terabytes of Oracle’s SAN Storage as part of its tiered storage infrastructure with Sun Storage Archive Manager software. Sun Storage Archive Manager is enabling Caltech to quickly access data in its archive and ensure that an archival tape copy is made using open standards for long-term preservation. It also provides a standard filesystem interface to a scalable hierarchical storage management system.
Furthermore, Caltech intends to leverage StorageTek Data Integrity Validation on the StorageTek T10000C tape drives to ensure there is no data loss during the transfer of files or over the course of time.
“Over the last 12 years, the LIGO project has generated three petabytes of data as a thousand scientists and engineers have worked to test Einstein’s theory of relativity and observe gravitational waves,” said Stuart Anderson, director of computing, for the LIGO Laboratory. “With the scope of the project and our intent to keep the data in perpetuity, a robust, scalable and open storage solution is absolutely vital. Oracle’s StorageTek tape storage has been central to our archive and its latest generation StorageTek T10000C tape drives have enabled us to manage more data and promises to provide additional data protection for our data, while preparing for significant data growth going forward.”