Wednesday, January 27, 2021

StrongBox signs with DKRZ

StrongBox Data Solutions, a leader in large scale data management, signed recently a very big deal in terms of managed volume of data.  The deal with DKRZ, the German climate computing centre, is super ambitious and will serve as a perfect data management case study for very large data environments.

DKRZ requirements are pretty impressive and clearly it doesn't exist lots of product able to manipulate and move the volume generated each day there. The project has 2 sides:
  • migration of 150PB of HPSS to LTFS,
  • workflow automation of 120PB of HPC data per year across storage tiers.
As the deal is signed for 5 years , it means a volume to manage of 600PB minimum.

StrongLink will offer the capability to consider multiple secondary storage targets. Among these multiple devices and generations, LTFS and S3 play a key role, giving transparent access to data via the middleware role of StrongLink wherever data reside.

The phase of migration from HPSS to LTFS marks a significant shift to agile metadata oriented workflows management.

StrongLink serves as an universal horizontal access layer via its global namespace capability across all tiers whatever is the access method such network file system and S3 for remote users. The product can also aggregate cloud-based storage if needed via back-end S3 API support.

The tape environment is designed with 5 tape libraries 90 tape drives and 75,000 cartridge slots representing on average 18 drives and 15,000 slots per library.

At scale, data movement is key to master cost. DKRZ uses a 3 tier model with a HPC tier 0, a 1PB tier 1 as cache with 30GB/s minimal throughput and a tape tier 2.

Such project shows one key capability of StrongLink with its metadata strong feature for new data but also extracting and harvesting ones from old or current data.

An other important aspect for DKRZ is the investment protection of current storage assets. Super deal for StrongBox, an other big deal is coming in Germany as well with 2PB of new data per day... impressive!!
Share:

0 commentaires: