Pure Storage Charts A Course To The Future Of Big Data

Pure Storage made a bundle of announcements at its 2nd annual Accelerate conference in San Francisco today.

It says it has added 25 new software features to its platforms that will further expand its appeal and fill in some gaps in its offerings.

The new ActiveCluster synchronous replication feature brings synchronous replication to the platform as a software upgrade in the Purity FA 5 release for the FlashArray product line. It will be freely available to all customers on supported FlashArray models, and provides an active-active cluster for two arrays provided they have an IP network with less than 5ms latency between them.

“People don’t want their life to suck just because they decided to protect their data,” said Tabriz Holz, Engineering Lead for ActiveCluster. She explained that customers don't really want to pay much more just to have synchronous replication, and don't want the hassle of a separately licensed feature. She called it a subscription to innovation which was a term in regular use by the company during the conference.

ActiveCluster fills a gap that has been much requested by European customers in particular, and should drive growth in this region.

Pure has also added lots of capability to its denser FlashBlade platform. FlashBlade will now scale up to 75 blades, providing up to 8 petabytes of storage space in a single namespace.

Purity FB 2, the operating system used by FlashBlade, will support VMware VVOLs using a highly available VASA provider that is hosted on the array. A VVOL on the FlashBlade array is just a volume, and can be mounted directly from the array, providing a useful mechanism for moving data into and out of VMware. Purity FA 5 will also support VVOLs.

Pure is also conscious that it needs to be an attractive choice for workloads both today and in the future. It is specifically aiming to provide features that new developers are looking for, like adding an S3 object interface to FlashBlade which implements a native object store inside the array.

FlashArray will also support a feature called Purity Run, which allows customers to run containers on the array itself. This is not an HCI play, and Pure explicitly called out that it will not be marketing the feature as HCI. It is designed for specific workloads that work best when co-located to the data on which they operate. I can imagine adding something like DataGravity via Purity Run to add extra analytics and insight to the Pure Storage based data platform.

Pure has updated its snapshot format to embed recovery metadata which makes the snapshot more portable. Snapshots can be made on FlashArray and then replicated to FlashBlade, and can now also be replicated to third-party NFS storage (*cough*NetApp*cough*). The failure/recovery capability of ActiveCluster uses the snapshot capability under the covers to catch up should the latency between arrays fall beyond the 5ms required for synchronous mode, providing automatic recovery from asynchronous mode back to synchronous mode. Pure has also introduced the CloudSnap feature that allows for these portable snapshots to be replicated to an S3 compatible cloud platform.

I asked CEO Scott Dietzen if the codebase for Purity was unified across the platforms, and he said that while they share a common heritage, Purity FA (for FlashArray) is different to Purity FB which was formerly called Elasticity. This will likely become more important as more features are added to Pure’s platforms due to the overhead of managing two disparate codebases, as well as the cloud-based Pure1 environment used for tracking customer telemetry data.

Pure is clearly plotting its path to be an attractive location for the huge amounts of data being generated by customers, much of which will not end up in the cloud. Dietzen pointed out that by 2020 over 50 zettabytes of data will be generated, much of it by sensors and other IoT style devices, but only 2.5 zettabytes of Internet bandwidth will exist. It will not be physically possible to move all of the data into the cloud, which means it will need to be stored somewhere for processing.

Pure Storage hopes to convince an increasing number of customers that the place to store their data is on Pure’s products, and is rapidly adding the functionality they require.

I traveled to San Francisco for Pure //Accelerate as a guest of Pure Storage.

This article first appeared in here.

See Also

Related items

Pure Storage Goes After NetApp’s Lunch With New FlashBlade

14 March 2016

Pure Storage launch a new product line, the FlashBlade, aimed at customers who want fast, dense, NFS addressable flash.

Pure Storage Posts Strong Q4 Numbers Post IPO

2 March 2016

Pure Storage has posted strong financial numbers in its first reported fiscal quarter since IPO.

SolidFire Aims To Make CapEx Great Again

7 June 2016

SolidFire want to make CapEx great again with its FlashForward Capacity Licensing model.