Storage tool assures CERN tape integrity

gallery

CERN, the European Organisation for Nuclear Research, is using a new tool to independently check the quality of large quantities of research data held on tape.

The new tool is helping the world's largest high-energy physics research centre assure important archived research remains accessible, while allowing its storage management team to quickly identify potential tape or data issues as well as keep up with migrating to the latest tape technologies.

Dr Charles Curran is a storage consultant and physicist at CERN, responsible for its tape storage operations. He told IT PRO how important tape is to the work CERN carries out as a low-cost storage medium, given the fact that the sheer scale of data produced does not easily lend itself to multiple backup copies.

"The sheer scale of the work involved means we can't afford to have offsite backup copies of data," he said, explaining why maintaining the integrity of original tape storage is so important.

"We also have to make sure the data is accessible and keep it organised, so we also have to make sure it is transferable to the next generation of tape media, but checking the quality of old tapes and verifying the integrity of the data they hold has historically made it difficult to keep up with the pace of migration."

CERN has recently installed the Tapewise Enterprise software tool from specialist Data Product Services (DPS) UK to ensure the quality of data held on tape, in the same way hard disk drive tools can perform disk bad block checking. This means that tape media problems and drive problems can be identified quickly and their scope found out before data is put at risk.

This also means the organisation is now better prepared to effectively manage new volumes of data created by its new Large Hadron Collider (LHC), which is scheduled to start operating in 2008. The data generated during the multi-year LHC project is expected to dwarf every other scientific experiment in history, amounting to 15 petabytes a year.

"We are already probably one of the biggest data holders in the world with eight petabytes of physics data held on 30,000 cartridges," said Curran. "But because of the LHC project, the number of tape cartridges we need will increase markedly."

As only tape has the capacity and cost to allow data to be kept semi-online inside a library, the primary LHC storage medium will be tape, using IBM 3584 and Sun StorageTek StreamLine libraries.

The use of Tapewise Enterprise will give CERN the confidence that the data can be read and written once and that the process of checking each tape can be automated to speed management and migration procedures.

The storage integrity problem is exacerbated as, each time there is a tape generation change, CERN migrates its data to take advantage of capacity and I/O speed increases, while maintaining an in-house storage management software system. As a result, Curran is also expecting a StorageTek Powderhorn library to reach its end-of-life in 2008, meaning a migration of thousands of tapes to a StreamLine replacement.

Using Tapewise, tape media management and data quality and reliability has improved, as has CERN's ability to more quickly identify if issues lie with its own software systems or with the tape manufacturers.

Email to a friend

Print this page