StorOne, Sanmina’s Viking Collaborate On High-Performance Storage

The new StorOne/Viking system is only available via TD Synnex, which is doing the integration for channel partners, says StorOne co-founder and CEO Gal Naor.

Software-defined storage technology developer StorOne Tuesday unveiled a partnership with Viking Enterprise Solutions under which its software and Viking servers will be integrated and sold by TD Synnex.

Gal Naor, co-founder and CEO of New York-based StorOne, touted the efficiency of the company’s storage software platform.

“Merely getting high performance is less of a challenge because you can throw more hardware in and you will get it,” Naor said. “The challenge is to provide the same solution to customers with a minimum amount of hardware. It means much more software, more algorithms and extra data protection while reducing complexity and not just adding more components or more third-party hardware.”

StorOne’s software allows storage systems to be built with industry-standard servers, with customers and partners able to choose any drive from any manufacturer and connect in any RAID configuration, Naor said.

“This is our technology,” he said. “We have more than 50 patents, all of them focused on how to bring customers the efficient solutions that they need.”

Viking Enterprise Solutions is a product division of San Jose, Calif.-based Sanmina, which is one of the world’s largest contract manufacturers developing products for a wide range of industries.

Working together, StorOne and Viking are building a storage system based on a Viking 2U chassis configured with AMD Rome or Milan processors, and 24 hard drives or flash drives, Naor said. Partners can also expand capacity with an external JBOD (just a bunch of disks) disk shelf. StorOne provides hybrid flash and hard-drive support to reduce costs, he said.

It also allows over 100,000 data snapshots for data protection, performance of more than 1 million IOPS with less than 1 millisecond latency, and sequential throughput of 33 GBps, he said.

The result is an enterprise-grade storage system that supports modern storage workloads including primary storage, backup, archiving, virtualization and cloud integration in a single platform, he said.

Naor cited as a potential workload a health-care customer that needed high performance for applications like storing PACS (picture archiving and communications system), MRIs or ultrasound images. The images could be automatically migrated on the fly via the NVMe protocol directly to the hard-drive tier of storage.

“This provides very low cost when the data is not in use even as the data is available to them all the time,” he said. “The data is not moving across different systems. There’s no deep archive that requires days to access it. It’s all on the same platform. It reduces complexity and is very easy to manage.”

The new StorOne/Viking system is only available via TD Synnex, which is doing the integration for channel partners, Naor said. StorOne only works via indirect channels, he said.

Looking forward, StorOne is focused on improving its storage software’s auto-tiering capabilities, Naor said.

“It's not enough just to move data from the upper tier to the lower tier,” he said. “The challenge is to do it without any configuration, all automatically. The system, we call it ‘AI Smart Auto Tiering,’ knows by itself how to increase or decrease the upper tier volume sizes in order to provide very high performance for data on the upper tier and then to move the data to the lower tier [when high performance is not needed].”

As a result, the data is stored via random access in the upper tier while it is being moved sequentially to the lower tier to increase performance, Naor said.

The improved auto tiering also learns data patterns so that reading of data for which performance is not needed, such as video surveillance data, is done directly from the lower data tier without first moving it to the faster tier, he said.

In another case, for AI learning, businesses need the performance of flash storage but can move the data to a hard-drive tier for longer-term storage to reduce the cost of AI, Naor said.