data.psc.edu — The Data Supercell

The PSC-developed Data Supercell is a low cost, high bandwidth, low latency, high reliability and high capacity disk-based data management solution. It provides academic, corporate, government and research partners with a convenient and affordable way to store and access their data, including extremely large data sets.

Storage Features

• Storage system architecture is based on file systems designed for high reliability and scalability, including data transfer verification and validation, providing unequaled durability for even the largest data sets.
• Basic DS service includes a single secure copy; DS system design allows a second copy to be easily added if required; future plans will allow further redundancy to be easily accommodated with a third copy replicated to our off-site storage.
• Files of any size can be stored, from smaller document collections or datasets to multi-terabyte backup sets. The Data Supercell is an ideal solution for off-site backup storage. Minimum storage allocation is 10TB.

Access Features

• Simple and secure interface methods make it easy to use for the typical user.
• Access methods are Grid FTP/Globus On-Line compatible.
• Interface methods can be customized when needed to provide flexible and expandable solutions to meet the needs of more demanding applications. Federated authentication available if required.

Better than Archiving

More cost-effective than tape

• Bandwidth ~ 24 times better than tape
• Latency ~10,000 times better than tape
• Scalable

Highly secure

Highest reliability for petascale storage

• Enhancements beyond standard RAID
• Options for geographical redundancy
• Optimize data replication and data movement

What was archive is now near-line

• Enable analysis of the largest datasets
  from gigabytes to terabytes to petabytes, over long periods.
• Preserve data models
  simplify software design and reduce maintenance costs.
• Maximize data utility
  Store metadata with data.
  Integrate structured, semi-structured, and unstructured data.
• Make the best decisions based on all available data
  Keep data useful, not buried in "write once, read never" archives.

In combination with Data-Intensive Computing (very large memory and analytics) the Data Supercell can provide exceptional business intelligence and results in record time.