DataDirect Networks (DDN) has launched a petabyte-scale scale-out NAS bundle, the GS7K, aimed at big data analytics use cases.
The GS7K is preconfigured with the IBM-developed General Parallel File System (GPFS) for scale-out NAS with parallel file system capability.
Optimising GPFS is a non-trivial task so DDN’s key selling point with the GS7K is that it takes the work out of setting up the parallel file system by bundling controller, file system and storage in a product.
Parallel file systems differ from traditional NAS file systems by being hugely scalable and by being able to use a single file system across many physical controller/storage nodes.
In a scale-out file system, controllers and/or storage can be added and the file system remains unchanged, but expanded or with added CPU power. This differs from traditional NAS systems that were limited to a single physical instance, often meaning islands of storage were created because capacity had been reached.
More on scale out NAS
- Scale-out NAS product survey: The big six
- A guide to scale-out NAS: The specialists and startups
- Definition: Scale-out storage
- Scale-out set to eclipse scale-up for enterprise storage
- How does a scale-out NAS environment affect storage management?
A base unit of the GS7K comprises a 4U unit that can house 16 SAS or SATA drives of 2TB, 4TB or 6TB capacity with the addition of flash drives possible; DDN recommends the use of 400GB flash drives. auto-tiering software ensures that hot data resides on the flash component.
A node can be expanded to 396TB with the addition of four 84-drive JBODs. A cluster can be built of more than 10 instances of GS7K head units plus JBODs to provide capacity in the petabytes. Throughput of 100Gbps is possible with 10 nodes; 11Gbps is possible from a single node.
The GS7K – which has Intel Ivy Bridge processors – is a re-working of the existing DDN GridScaler scale-out NAS product, which is supplied with DDN SFA 12K or 7K storage back end.
Traditional NAS operations can be run on the GS7K as NFS file system access is also provided.
The GS7K is aimed at big data analytics and high-performance computing use cases such as life sciences, financial services and petrochemicals, as well as large-scale file access workloads such as media and entertainment.