ActiveStor Flash and ActiveStor Ultra XL target I/O profiles from rapid AI/ML to processing-heavy modelling but remain firmly on-prem, with object and cloud declared interesting
- Antony Adshead,Storage Editor
Published: 09 Sep 2022 9: 29
Panasas has broadened its scale-out NAS offer to encompass high-performance and capacity options with general option of ActiveStor Flash and ActiveStor Ultra XL. Both products aim at a variety of workloads with regards to quality and I/O profile that fall within the high-performance computing (HPC) to artificial intelligence/machine learning (AI/ML) continuum.
Speaking with ComputerWeekly.com, the business also revealed the limits of its fascination with object storage, along with its applying for grants cloud storage, where it does not have any presence currently.
The Panasas ActiveStor systems have already been tailored to a variety of workloads, that may mean file storage profiles that go from many, many really small files to an inferior number of large ones.
ActiveStor Flash is really a solely NVMe flash-based hardware appliance targeted at smaller file sizes where rapid access is necessary. Its ASF-100 nodes are 4U form factor and use up to 3.84TB of M.2 and 46TB of U.2 NVMe. DRAM and NVDIMM offer faster cache-level method of storing working data.
Meanwhile, ActiveStor Ultra XL is targeted at larger capacities and bigger file sizes. An ASU-100XL node runs to 160TB but quadruple that for minimum configuration mostly made up of spinning disk HDD and several faster M.2 NVMe capacity.
Both systems, both running PanFS, have benefited from controller OS and file system upgrades in version 9.2 that enable the client to deploy storage blades under an individual namespace. But with volumes intended to suit workloads of differing I/O characteristics so, smaller and fast, or cooler and larger under a unitary pane of glass, said Curtis Anderson, software architect at Panasas.
He added: We were a one-platform company until May. Then we’d two new platforms which are designed on the capability to use multiple media types, with metadata likely to NVMe for instance, SSDs for small files around 1.5MB and HDD for large files.
The Panasas name for the functionality is Dynamic Disk Acceleration, that is the automated direction to different tiers of storage.
The reason behind the shift? The problem was, imagine if a person is running HPC and really wants to run another workload? said Anderson.
The enhancements to PanFS enable that and the engineering behind it had been, said Anderson, a moderately- sized lift that involved refactoring PanFS to take care of new hardware types also to select and qualify those products for use with the machine.
But think about object storage, considering that so much unstructured data Panasass bread and butter is currently in object storage format?
Anderson said: Panasas is made as a Posix file system but along with an object store, that was produced by 1999, so before Amazons S3. It gets the characteristics of scaling and growth, etc, that object storage has, but we dont offer access. It works differently to S3.
Marketing and products VP Jeff Whitaker added: Object storage is of interest, however when it involves how the the greater part of individuals access data, its file-based. The development side of AI/ML often happens in the cloud, however, so its definitely something were thinking about as we progress.
In a context where in fact the cloud is now increasingly important and several suppliers provide possibility to store data in the cloud, what’s the Panasas strategy here?
The business is firmly still in the on-prem hardware camp but, much like object storage, it really is considering possibilities, said Whitaker. At this time, we have been an appliance-based datacentre platform, not software-only, and from what weve observed in the marketplace, 85-90% of the marketplace continues to be on-prem.
He added: Customers battle to get performance from cloud-based storage. Cloud providers need to throttle storage so their networks arent saturated. Absolutely, customers are moving to the cloud and doing more there, so we have been considering different scenarios and handling S3, with partners.
Read more on SAN, NAS, solid state, RAID
Panasass 10x saving bumps NetApp from HPC storage deployment
Panasas on data architecture – filesystem finessein the CS Wild West
Panasas on data architecture – haystacks & hammers, choose your poison
Panasas storage yearns for expanded enterprise presence