At its Accelerate conference in Austin Tuesday, Pure Storage introduced new products meant to make it easier for enterprises to switch to hybrid cloud environments and use machine learning technologies.
An all-flash storage pioneer, the company also announced a mix of new storage hardware, including a more affordable all-flash array for less mission-critical needs, such as data backup.
The announcements at Pure’s Accelerate 2019 included:
- The availability of Cloud Block Store for Amazon Web Services, which allows customers to move their applications and data to the cloud while using Pure’s software.
- AI Data Hub, which allows organizations to quickly deploy AI workloads by unifying data that is historically siloed in data warehouses, data lakes, AI clusters, or software development platforms.
- FlashArray C, a cost-effective all-flash array for tier-two data. It provides the same flash reliability but adds 2 to 4 milliseconds of latency to make it more affordable. (You can get sub-millisecond latency with some faster flash technologies.)
Pure, which competes against the likes of Dell EMC, Hitachi, Hewlett Packard Enterprise, IBM, and NetApp in the all-flash storage market, hopes to grab a bigger share of the lucrative data protection and recovery market with its FlashArray C announcement, analysts say.
As prices for solid state drives (SSDs) continue to drop, Pure and its competitors can target enterprises that need to upgrade its data backup solutions, said Eric Burgener, research VP with IDC’s Infrastructure Systems, Platforms, and Technologies group. Two years ago, HPE’s Nimble Storage announced its own flash array for secondary storage.
Pure saw the potential for selling all-flash storage for secondary workloads when it introduced a version of its FlashBlade storage technology and saw some companies start to use it for data backup, even though it wasn’t designed for that purpose, Burgener said.
“There is a total cost of ownership story for using all-flash for secondary workloads,” he said. “It can move data faster. It uses less power. It’s more reliable and makes better use of floor space – and it’s only slightly more expensive than HDD (hard disk drive),” he said.
Pure Storage Goes Cloud
One of the two new Pure cloud offerings, Cloud Block Store for AWS, aims at companies building hybrid cloud environments and moving applications and data to public clouds or using public clouds for disaster recovery or testing and development, Chadd Kenney, Pure’s VP of products and solutions, explained.
“It has the same APIs as our on-premise solution, so any workflows or automation you’ve built works great,” he said in an interview with Data Center Knowledge. “It also has the same efficiency software – deduplication, compression, and thin provisioning – to reduce the capacity footprint in the cloud.”
Pure’s existing Evergreen Storage Service customers can use the same contracts to take advantage of Cloud Block Store, Kenney said. Evergreen is Pure’s pay-as-you-go on-premises storage service. A customer who’s installed 500TB through Evergreen, for example, can move half of that capacity to Cloud Block Store for no extra charge.
“With most solutions, you have to buy 250TB of new storage in the cloud and then waste that space on-premise, but with us, you can transfer that subscription to the public cloud and have 250TB on-premise and 250TB in the public cloud,” Kenney said. “That gives you investment protection.”
The second new cloud offering Pure introduced Tuesday was CloudSnap for Microsoft Azure, a cloud-based backup solution. It’s been offering the same service for AWS.
AI Data Hub
The AI Data Hub is the latest product to come out of Pure’s partnership with Nvidia. The two companies previously released AIRI, which stands for AI-Ready Infrastructure and includes Nvidia’s DGX-1 supercomputer hardware and Pure’s FlashBlade for storage.
AI Data Hub is a reference architecture that uses Kubernetes for orchestration across the data pipeline, Amy Flower, VP of strategy and solutions in Pure’s FlashBlade Business Unit, told us.
It creates an infrastructure to quickly ingest data, clean and label it, and then enable data scientists to use subsets of data to build, train, and deploy machine-learning models. “Unifying the silos is the big objective here,” Flower said.
In related news, the company announced availability of AIRI-Ready Data Centers, a group of service providers that will offer AI infrastructure through a Software-as-a-Service model using AIRI. Among those partners is Core Scientific, the company said.
Furthermore, Pure announced availability of DirectMemory Modules, which provide faster array performance using Pure’s DirectMemory Cache software and the new Intel Optane storage-class memory. The modules can be plugged directly into Pure’s FlashArray X70 and X90 hardware, the company said.
Pure also announced a new FlashBlade with 150 blades. According to Flower, it is twice the size of the previous version.