The price convergence of hard disk and flash storage has been the topic of much discussion in recent years, as organizations seek more efficient solutions for ever increasing amounts of data. Today Pure Storage believes that it has solved the condundrum for enterprises that need to store large amounts of unstructured data, claiming that its new FlashBlade//E data repository can beat out hard disk alternatives on all value metrics – including price.
Historically, buyers have filled their data centers with hard disk solutions because they are cheaper to run and manage, compromising on efficiency and performance. Meanwhile, flash-based solutions have typically been considered superior, but often fall short in the buying process due to price concerns at scale. Pure Storage is hoping that its FlashBlade//E offering for unstructured data can eliminate that concern.
And Pure isn’t talking about a cost point that has to take into account total cost of ownership ( ongoing resources, management etc.). Pure Storage is claiming that its FlashBlade//E solution will be offered at under $0.20 per GB including three years of service, before taking into account other total cost of ownership factors. However, the platform starts at 4 petabytes and scales from there (so is for large unstructured data workloads).
Cost isn’t the only differentiator. Pure is also stating that FlashBlade//E has sustainable performance at its core, consuming up to 5x less power than the disk-based systems it seeks to replace (and of course, the larger the system, the greater the efficiency).
In addition, FlashBlade//E has 10 to 20 times more reliability than hard disk based systems, claims Pure.
We got the chance to sit down with Patrick Smith, Pure Storage EMEA CTO, to talk through the significance of the announcement. Smith says:
Today for very large data use cases, flash storage has been considered to be too expensive for those workloads. One of the things that we at Pure have always been targeting, ever since the foundation of Pure in 2009, is to be able to compete with disk based storage for different workloads.
We started off by going after the world of database and virtualization – we’ve now made that mainstream. And now we’re going after other areas that disk is considered to have a stronghold.
So what we’re announcing is a new model within the FlashBlade portfolio, an unstructured data platform that brings new levels of capacity, affordability and environmental footprint to the data center.
Smith says that for a long time the market has spoken about flash and hard disk price convergence, or the point at which flash becomes cheaper than hard disk drives – given that one of the barrier to organizations deploying flash at scale has been cost.
The reason that Pure Storage has been able to overcome this, adds Smith, is because of its decision back in 2014 to build its own flash drives. It moved away from using commercial off the shelf SSDs and decided it wanted to utilize its own software capabilities on its own hardware – creating its own flash modules. This has allowed it to control, and accelerate, the capacity of flash storage ahead of other SSD manufacturers, says Smith. He adds:
When we talk about 20 cents per GB, that is simply the acquisition cost and maintenance cost for three years. So then, operations brings benefits beyond the acquisition cost in terms of the lowest level of operational overhead that we can deliver, that actually any storage provider can deliver.
We believe that coupled with the benefits of simplicity of operation, of lower data centre footprint, lower power consumption, will see an approximately 40% lower TCO [total cost of ownership] over six years, which then is a really significant benefit to our customers.
The use case
It’s worth reiterating that the FlashBlade//E offering is specifically catered to highly scalable unstructured data workloads. In terms of what these look like, Smith explains:
If we look at the use cases that we expect to see customers deploying FlashBlade//E, they’re really very broad across the unstructured data environment. We’re increasingly seeing people wanting to store large amounts of content for longer, and not all of that content may be needed in a dynamic manner, so they need very large content stores – in the petabyte region.
The same is true of data protection repositories, where organizations want to be able to complement a high performance data protection repository with a long term data retention environment. FlashBlade//E is the perfect location for that long-term data protection repository.
We’re also seeing, in the world of data analytics, data lakes growing larger and larger as customers want to store datasets for longer and longer periods. The value of data and the desire to not want to remove data means that these data lakes are growing extensively and being able to deliver that with an efficient platform that can still support active workloads, I think absolutely essential.
An explosion of unstructured data
Pure Storage says that for global organizations, unstructured data capacity is expected to grow by 10x before 2030 – adding that for large capacity, price censitive workloads that currently use disk- based storage solutions, this growth is unsustainable. Smith says:
What’s interesting about the majority of that unstructured data is that it doesn’t require high performance. By the nature of it not requiring high performance, most of that unstructured data sits on spinning hard drives today.
Those spinning hard drives are the worst contributors to data center floor space utilization, to data centre power utilization and to greenhouse gas emissions, as we look at the storage fleet in the data centre. And so being able to meet this massive growth in demand for unstructured data isn’t sustainable with today’s platforms in our view.
As noted above, beyond the cost factors, Pure Storage is keenly aware of organizations’ sustainability priorities over the coming years.
A recent survey conducted by Pure, which fielded responses from 1,000 sustainability program directors across the US, UK, France and Germany, found that most (78%) said that their company’s leadership is treating sustainability initiatives as a priority. However, whilst the majority said that they plan to meet sustainability goals within three to seven years (56%), only half (51%) of those surveyed said that they are on track to meet their goals.
And whilst 86% of sustainability program managers agree that companies cannot reach their sustainability goals without significantly reducing their technology infrastructure energy usage, some 81% predict the impact of this infrastructure on a company’s carbon footprint will increase in the next 12 months.
Pure – and its all-flash offering – aim to help organizations close this gap between sustainability priorities and action on the ground. Smith says:
Where we see today’s disk platforms under-delivering…firstly, too much space in the data center. If we look at most modern data centers, even data centers being built today, they are struggling to cope with the space and power demands of disk based storage. People are trying to balance more compute and more data in the same data centre.
But the other challenge is too much energy. As you’ll have seen in so many press announcements, there’s a desire to reduce energy utilization, both in terms of cost, but also the ability to get access to energy for data centers. That 10x growth in footprint is going to cause real problems in terms of energy consumption, if we don’t change the dynamic in the data center.
But people power is a concern too. Smith adds:
And then finally, too many resources, whether those resources are racks, energy, e-waste, or resources to manage the environment. If that data center footprint to support unstructured data grows by 10x organizations simply don’t have the appetite or the budgets to increase their employee base by 10x, to support that growing environment. So we have to do something about the challenges that today’s data platforms for this unstructured data presents to organizations.
Data requirements for organizations are only going to increase over the coming years – and the platforms that support data initiatives are going to need to be increasingly efficient and able to support complex environments. Data centers can’t keep relying on hard disk as these environments scale; not just from a cost perspective, but an environmental concern perspective too. The validation for Pure’s claims will of course be in the uptake and customer testimonials – but from the confidence of the team’s pitch, and its track record, it seems Pure’s all-flash future is becoming a much closer reality.