Durham DiRAC gets cosmological data on tape

The DiRAC supercomputing facility at Durham University in the UK is using two Spectra tape libraries to store cosmological data about galactic evolution.

DiRAC (Distributed Research using Advanced Computing) is a federated facility, funded by the UK’s Science and Technologies Facilities Council (STFC). It’s spread across four sites at Cambridge, Durham, Edinburgh and Leicester.  Durham houses the DiRAC Memory Intensive Service in its Institute for Computational Cosmology.

Durham University’s Ogden Centre for Fundamental Physics houses the ICC and DiRAC.

This Durham DiRAC high-performance computing system is used for theoretical modeling and HPC-based research in particle physics, astronomy and cosmology, and nuclear physics. The Durham HPC cluster has 812 nodes, 58,700 compute cores, and 230TB of DRAM backed up with 10PB of primary storage. It uses the GPFS and Lustre parallel file systems.

Spectra T950 with Terapack visible.

It’s anticipated that DiRAC will generate 20PB of stored data by 2022. The data chronicles galactic formation and evolution, and will need to be accessible for 15 years. DiRAC staff are using one Spectra T950 tape library to store this data, with a second T950 purchased as part of a general DiRAC phase-3 upgrade.

Both have four LTO-8 drives and media. Data is accessed using the LTFS protocol. Atempo’s Miria software is the selected data mover, managing both backup and archive processes from the primary storage.

The LTFS (Linear Tape File System) presents a tape drive that writes files sequentially as if it were a disk-based file:folder device, with files listed for random and not sequential access. There is an index (directory) stored on a tape cartridge which lists the cartridge’s contents as if they were in a file:folder structure. LTFS software in a server accessing a tape library reads a cartridge’s index and presents it in a file:folder format on screen.

The T950 scales from 50 to 10,250 LTO slots, with loading via ten-cartridge TeraPack trays, and has up to 120 tape drives. 

Read a case study here.

Bootnote: DiRAC is a pun on the name of Paul Dirac, a highly-regarded English physicist involved with quantum mechanics and quantum electrodynamics.