Pure Storage unifies flash block and file experience to reduce management complexity and drive efficiencies

(© Sergey Nivens - Shutterstock)

In a world where data management is becoming increasingly complex for organizations, and where data use is soaring, Pure Storage is seeking to simplify access and management for customers across its FlashArray product line. This has been a long running priority for Pure, but its latest development will see customers get access to a unified block and file experience across ‘global storage pools’, eliminating the need for legacy multi-protocol add-ons that make working with and accessing data difficult. 

Pure has said that the new global storage pools, alongside unified policy management and VM-aware storage management capabilities, means that customers using FlashArray will be able to lower their storage total cost of ownership by up to 58% (source: Enterprise Strategy Group). 

Speaking with Patrick Smith, Pure Storage EMEA CTO, ahead of the announcement, he explains: 

I think it's a really interesting area, because lots of our customers talk about our file capabilities on FlashArray and they say: when will we have a product to truly compete with our incumbent file platforms that we've had for 20 years? And to be honest, they don’t like those platforms. 

They want all the benefits that they have with Pure on FlashArray and the block environment, and bring it to the world of file. So we have been putting a huge amount of effort into this and making sure we can deliver file services in what I call ‘the Pure way’. 

Making sure that it’s simple, making sure that it’s efficient.

Legacy multi-protocol block and file storage solutions have been around for decades. However, according to Pure, these products failed to deliver the flexibility and efficiency that unified arrays promised. 

Built for either block, or file, these protocols were bolted on and increased management complexity at scale. Pure’s approach aims to unify this storage experience and deliver an architecture where both block and file are both native - or as Smith says, treated as “first class citizens”. He explains: 

As we've built file for FlashArray, one of the things we've been really conscious to do is to make sure that file isn't a bolt on, on top of the FlashArray block capabilities. We really wanted to make sure that file on FlashArray is treated as a first class citizen, alongside block.

That gives us the full set of capabilities - and they all sit natively on the storage pool that is FlashArray. And it gives us lots of capabilities that you don't have if you add file as an afterthought. 

So by making it a first class citizen, you get all of the benefits of things like deduplication of all the data on the system, rather than having discrete pools of information where there's no commonality and you can't deduplicate up across those different pools. 

It also gives us a really big benefit in terms of manageability of that environment, because you don't have to manage these environments entirely separately, they've got the same management paradigm. 

It's all managed through our management platform, Pure1. So it significantly reduces the management overhead compared to platforms offered by our competitors. 

With the FlashArray Unified Block and File Platform, Pure is offering customers: 

• Global Storage Pools - a global storage pool that aims to eliminate the complexity associated with data growth on legacy unified arrays. Before, according to Pure, administrators needed to pre-plan every storage update and request. It says that now they can simply use what they need, across block and file, with a non-disruptive expansion on the fly and unlimited file system sizes. 

• Unified Policy Management - Pure Storage says that it has now eliminated the multi-layered management required by legacy storage platforms, giving administrators precise management of the specific storage service they want to deploy and control. Pure says that by unifying policy management, operations can be learned quickly and applied to everything, both block and file.

• VM-Aware Storage Capabilities - Pure Storage has also introduced VM-Aware Storage, giving customers deeper visibility at the granularity of the virtual machine. The granular visibility and management Pure is bringing to file is also available to VMs with VM-aware Storage. Administrators can natively manage VMs on FlashArray, including VM-level statistics, snapshots, quotas, and policies.   

VMWare granularity 

Smith is also keen to highlight the significance of the VMware storage capabilities that are being introduced as part of the file and block unified FlashArray. He says that the push to certify FlashArray file with VMware is because a lot of the vendor’s customers that have gone down the file route, using it to support VMware. He explains:  

What we've built is effectively a storage pool that can be, and is the size of ,the whole array. And then all of the VMs, plus all of the block devices, all sit in that storage pool. So, a VM can grow to the size of the whole array, if need be, with no moving of that VM’s data store around. So really maximum flexibility, as we look at how it's run. 

And that then leads us into this concept of VM granularity. So if you look at how legacy file platforms have adopted VMWare - you have a storage pool, you create a file system on top of that storage pool, and then within that file system, you create your VM data stores, and you put lots of VM data stores within that file system. 

The challenge then comes, if you want to do any operations on a VM data store, you need to do it at the file system level. The most common one, especially today with the rise of ransomware, is people using snapshots as a restore mechanism. If you want to take a snapshot, you have to take a snapshot of the whole file system. And then if you want to restore from that snapshot, you have to restore the whole file system snapshots, which may include tens, or even hundreds of VM data stores. 

Smith says that this scenario becomes costly, complex and unmanageable for customers. As such, Pure is bringing managed directories to the file unified environment. Smith says: 

So you end up with this bloating, based on operating at a file system level. We looked at that and we decided that there was a better way to manage VM data stores on our file solution. 

So we've come up with a concept of managed directories. And managed directories allow the administrator to create one directory within the overall file platform, per VM. But then the benefit of a managed directory is that you can manage at a directory level - that may sound like a really dumb thing to say, but it means getting things like statistics at a VM level, simply by looking at the manage directory statistics. 

But really importantly, it means being able to take snapshots at a managed directory level. And so suddenly, you can see that you've got one managed directory per VM. And you can take snapshots and restore snapshots at a VM level granularity. You can also apply quotas and policies at that managed directory level. 

So suddenly, rather than managing your VMware environment at a file system level. You've taken it to a VM level granularity which is really pretty unique in this space. 

My take

The technicalities of this are significant, but what’s important to remember is that this is aimed at reducing data management complexity. In a world where enterprises are grappling with data at scale, it's these fundamentals that will enable broader insights and transformations. We will be at Pure’s user conference in Las Vegas in June, where we look forward to speaking to more customers to dive deeper into what’s possible as a result.