Seagate offers California Uni genome data storage K-drives

UCSC becoming part of application developers ecosystem

Seagate has given the University of California, Santa Cruz a petabyte of Kinetic drives to store and access genomic data.

Kinetic disk drives have direct access over the Ethernet, and a key:value store structure with an object-style PUT and GET interface.

Andy Hospodor, UCSC’s executive director of the Storage Systems Research Centre (SSRC), part of the institution's Jack Baskin School of Engineering, said: “This gift provides the basis for a major research program on storage of genomic data."

The promised benefits include savings on hardware and software complexity compared with existing disk IO stacks.

There needs to be an ecosystem of application developers writing software that uses these drives for them to become popular, and Seagate is trying to kick-start such an ecosystem into existence.

We know Seagate is working with Europe’s atom-splitting CERN facility to do this. It’s also set up a Kinetic Open Storage Project with Toshiba, Western Digital, Cisco, Cleversafe (now IBM), Dell, DigitalSense, NetApp, Red Hat and Scality. The aim is to promote object storage on disk drives.

The disk drives can't be used in traditional drive arrays and require system/application software additions to manage data writing and reading instead of existing disk drive IO access software stacks.

The School of Engineering received this gift, 1PB of kinetic drives and 1.5PB of SATA disk drives. As far as we know K-drives are either 2TB in capacity or 8TB with shingled recording; so that means up to 250 4TB disks, or 125 8TB ones, or somewhere in the middle.

The SATA drives will go into existing genomic data storage clusters.

UCSC will use the K-drives to explore ways of organising access to large-scale genomic data, using object stores and proposed open-source standards (APIs) for genomic data being developed by the Global Alliance for Genomics and Health. However, it will also look at other ways of using its bunch of K-drives.

One idea is to process big data with storage local to the servers instead of shipping it to them from external arrays.

We’re thinking K-drive DAS in a server cluster with control plane software looking after data placement and access on the drives.

Possibly, Seagate is thinking about having its in-house EVault cloud backup and archive storage business use Kinetic drives as well. It currently has a deal with Iron Mountain for it to resell these services with Seagate kit stored in Iron Mountain’s underground vault in western Pennsylvania. ®