Panasas' dual NAS offering targets multiple analytical workloads

Panasas Has expanded its scale-out NAS product range to include high performance and capacity options and general availability of ActiveStor Flash and ActiveStor Ultra XL. Both products target a range of workloads in terms of file size and I/O profile that fall on the high-performance computing (HPC) to artificial intelligence/machine learning (AI/ML) continuum.

While speaking with ComputerWeekly.com, the company also revealed The limitations of its interest in object storage, as well as its thoughts on cloud storage, are not there yet.

Panasas ActiveStor systems have been targeted for a range of workloads Customized, which means that the file storage profile can go from many very small files to a few very large files.

ActiveStor Flash is a hardware device based entirely on NVMe flash memory, For smaller files that need quick access. Its ASF-100 nodes come in a 4U form factor and can take up to 3.84TB of M.2 and 46TB of U.2 NVMe. DRAM and NVDIMM provide a faster cache-level method of storing working data.

Meanwhile, ActiveStor Ultra XL targets greater capacity and larger file size. An ASU-100XL node runs up to 160TB – but four times that for the minimum configuration – mostly including spinning disk HDDs and some faster M.2 NVMe capacity.

Both systems running PanFS benefit from the 9.2 release The controller OS and file system upgrades allow customers to deploy storage blade namespaces under a single system. “But volumes are created to accommodate workloads with different I/O characteristics — thus, smaller and faster, or cooler and larger — under a single management platform,” said Curtis Anderson, software architect at Panasas. he added: “Until May, we were a single platform company. Then we had two new platforms built using multiple Capability for media types such as metadata to NVMe, SSD for small files up to 1.5MB, HDD for large files.”

The Panasas name for this feature is Dynamic Disk Acceleration, which is an automatic direction to different storage tiers. The reason for the shift? “The question is, what if a customer is running HPC and wants to run another workload?” Anderson said. Behind this, Anderson said, is a “mid-scale boost” that involves refactoring PanFS to handle new hardware types, and selecting and qualifying those products for use with the system.

But considering so much unstructured data – Panasas’ Livelihood – It’s all object storage formats now, so what about object storage?

Anderson said: “Panasas is built as a Posix filesystem, But at the top level object storage, it was developed in 1999, so before Amazon’s S3. It has the scalability and growth properties that object storage has, but we don’t provide access. The way it works with S3 Different.”

Jeff Whitaker, VP of Marketing and Product Adding: “Object storage is interesting, but when it comes to how the vast majority of people access data, it’s file-based. However, the development side of AI/ML often happens in the cloud, so as we go forward, It’s definitely something we’re interested in.” What is Panasas’ strategy in a context where the cloud is becoming more important and many vendors offer the possibility to store data in the cloud?

The company is still firmly in the local hardware camp, but Object storage is looking for possibilities, Whitaker said. “Right now, we’re an appliance-based data center platform, not just software, and from what we’ve seen in the market, 85-90% of the market is still on-premises.”

He added: “It is difficult for customers to get performance from cloud-based storage. Cloud providers Storage has to be limited so they don’t saturate their network. Of course, customers are moving to the cloud and doing more things there, so we’re looking at different scenarios and working with partners on S3.”

Read more about SAN, NAS, Solid State, RAID

10x savings on Panasas let NetApp deploy from HPC storage Author: Anthony Ad Head

Panasas on Data Architecture – Haystacks and Hammers, Choose Your Poison

go through: Adrian Bridgewater