How Pure Storage intends to continue disrupting data, databases, analytics

by News Desk   8 February, 2017
How Pure Storage intends to continue disrupting data, databases, analytics

Matt Kixmoeller is Vice President Products at Pure Storage.

When we started Pure Storage in 2009, we laid-out a bold but simple plan. We saw the coming of flash technology as an opportunity to re-think storage entirely, and we used flash as an opportunity to build a new class of smart storage solutions, focusing our innovation in three key areas:

#1 Build affordable all-flash storage

#2 Make storage simple and self-driving

#3 Make storage evergreen

Seven years later, our original Pure Storage mission is well on the path to being accomplished. We are proud to have had an outsized impact on the $35 billion storage market, as the storage behemoths have all pivoted to chase Pure Storage, attempting to emulate both our product strategy and business model.

As we look at the next chapter for Pure Storage and for the industry, we are focused on a central theme: exponentially expanding the value of data. We aim to deliver a data platform that enables our customers to build a new class of applications and to extract new insights from data – we call this data advantage.

With data advantage as our focus, we have set four additional new innovation priorities for the next five years at Pure Storage:

#4 Bring flash to large-scale unstructured data

#5 Transform application data storage again with Non-Volatile Memory Express

#6 Deliver next-generation converged infrastructure, optimised for cloud and new stack applications

#7 Make application development and data protection effortless in the cloud era

To be clear, we are far from done with our first three innovations areas! The storage industry is still only about 15% into the all-flash transition, and we are continuing to innovate with SaaS-based management, predictive analytics, and even more business model transformation.

As the businesses of legacy storage companies continue to falter, some have asked why build a storage company at all? Our belief is simple: in the increasingly digital economy data is key.

Yet we are just at the beginning of leveraging machine learning and advanced analytics to take data insights beyond even human prediction. Differentiation lies in having the most data, analysing it the fastest, discovering new insights, and ultimately delivering new products and experiences not possible without it.

Yet in most datacentres today, data is also the key bottleneck. Data on disk is the last mechanical gate in the compute chain. Data is where compromises are ultimately made: how much to store, how much can be analysed, how long to wait for answers, compromises no one wants to make.

The IT landscape today is dominated by vendors who are treating data as a compromise. Legacy storage vendors are retrofitting flash to decades-old systems, perpetuating the status quo versus embracing the true potential of flash. Public cloud vendors are making it cheap to store data but expensive to use it. Hyper-converged vendors are focusing on reducing cost by making storage just an integrated commodity function versus a platform for strategic advantage.

These vendors are all pivoting to low cost and low differentiation data strategies, versus investing in a platform to drive data advantage. Pure Storage believes there is a place in the market for a pure-play data innovator, wholly-focused on driving data advantage.

#4 Bring flash to large-scale unstructured data

Pure Storage’s success has been built upon helping core business applications and clouds that run them to be faster and more efficient. But this mission ignored the majority of data, the vast pools of unstructured data that in the world of digital sensors and machine generated everything is being created faster than it can be analysed.

Earlier this year we introduced FlashBlade, the second major product from Pure Storage, aimed squarely at tackling the big data problems and opportunities of tomorrow.

#5 Transform application data storage again with Non-Volatile Memory Express

Our founding vision for Pure Storage was to eradicate the spinning disk from the datacentre, but it turns out that the remnants of disk remain in the storage architecture even after the disk is gone. Legacy protocols, most notably SCSI, remain the archaic way we still speak to both flash within storage arrays, and to storage arrays over the storage network. It is time to banish SCSI for good.

The advent of the new Non-Volatile Memory Express and Non-Volatile Memory Express/F protocols hold the promise to do just that. Non-Volatile Memory Express over PCIe enables storage arrays to eliminate serial-attached SCSI links from their controller processors to flash. Non-Volatile Memory Express/F enables dramatically faster, SCSI-less connections between servers and storage arrays, making them appear local to servers and enabling local storage performance with networked storage shared efficiency and manageability.

Non-Volatile Memory Express and Non-Volatile Memory Express/F have huge potential to open-up the next generation of advances in storage-centric computing, but they face the same challenges that flash did a decade ago: cost, complexity, and interoperability. Early attempts at Non-Volatile Memory Express are once again focused at the ultra-high-end of the market, limiting its market potential.

Pure Storage’s approach to Non-Volatile Memory Express is simple, and perhaps unsurprisingly similar to how we first approached flash. We believe Non-Volatile Memory Express is for everyone. We are building a set of Non-Volatile Memory Express-enabled products that will allow mainstream Non-Volatile Memory Express adoption across a wide set of use cases, not just for the performance elite.

We believe that Non-Volatile Memory Express solutions should incorporate all the critical data management features that customers need and expect, HA, RAID, data reduction, encryption, snapshots, and replication. We believe Non-Volatile Memory Express should be an Evergreen evolution.

The move to Non-Volatile Memory Express will be a multi-year transition, and we are committed to taking our customers there in a non-disruptive, investment-protected, Evergreen fashion.

We believe Non-Volatile Memory Express will grow the market for shared storage. Non-Volatile Memory Express/F and Non-Volatile Memory Express enable these applications to leverage what looks like server-local storage, but to now take advantage of shared storage resiliency, efficiency, and manageability.

 

Pure Storage’s approach to Non-Volatile Memory Express is simple, and perhaps unsurprisingly similar to how we first approached flash. We believe Non-Volatile Memory Express is for everyone. (Graphic NVMexpress.org)

 

Non-Volatile Memory Express and Non-Volatile Memory Express/F have huge potential to open-up the next generation of advances in storage-centric computing, but they face the same challenges that flash did a decade ago: cost, complexity, and interoperability. (Graphic NVMexpress.org)

#6 Deliver next-generation converged infrastructure, optimised for cloud and new stack applications

Storage does not exist as an island, it is deployed with compute, networking, virtualisation, and all this is ultimately in the service of applications and users. Nearly a decade ago the concept of converged infrastructure was born to gain the advantages of an end-to-end tested stack. We see an opportunity to re-think converged infrastructure solutions for the modern cloud era, an era that demands greater agility and uses public cloud as the key comparison yardstick.

It was this need to modernise that first inspired us to create FlashStack converged infrastructure solutions over two years ago. Pure Storage and Cisco Systems recently announced an expansion of collaboration around FlashStack. We believe FlashStack is the most reliable, powerful, and agile private cloud foundation available.

Looking further beyond, we see an opportunity to redefine the lines between databases, analytic stores, and data storage. In the end, they are all about interacting with your data. Should they not be much more seamless? Can we blur the lines to create a new data tier to power next-generation applications?

#7 Make application development and data protection effortless in the cloud era

Delivering reliable and efficient storage is only a very small piece of the data lifecycle. While storage has improved, data management processes have remained largely unchanged for decades, and drive much of the proliferation of unnecessary data copies in the enterprise. It is also a whole new era, a cloud era, where customers are increasingly looking to leverage the public cloud to both develop applications and to retain data for long-term protection.

So, we ask ourselves a set of what if questions:

  • What if application developers could have access to unlimited, instant, and space-efficient copies of their data?
  • What if they could control and manage these copies without burdening a storage administrator?
  • What if these data copies could be automated as part of a DevOps cycle?
  • What if your storage array and your next-generation converged infrastructure was just another piece of infrastructure that acted as code in your environment?
  • What if on-premises private cloud infrastructure could be seamlessly integrated with the public cloud?
  • What if data could be easily moved between on-premises and the cloud, enabling development anywhere, deployment anywhere, and burst capacity anywhere?
  • What if this data movement could be orchestrated end-to-end?
  • Delivered as a SaaS-managed experience, with custom-tailored views for everyone involved in the process, enabling self-service and agility?
  • And what if all this was just as easy as Pure Storage is today, and just worked?

We see the opportunity to link on-premises storage arrays seamlessly with the public cloud, integrating application development, DevOps, and data protection workflows seamlessly with the storage. This enables speed and agility in application development, and allows customers to shift their workloads between private and public clouds as it suits their needs.

So, what does it all add up to?

We believe that the potential of data is so huge, so integral to the future of the digital economy, that the market deserves a pure-play company that focuses wholly on delivering data advantage.

  • Big data
  • Fast data
  • Data analytics
  • Data science and machine learning
  • Data orchestration between clouds, and data lifecycle management

(Image Shutterstock)

Key takeaways

  • Customers are looking to leverage public cloud to both develop applications and to retain data for long-term protection
  • Data on disk is the last mechanical gate in the compute chain
  • Delivering reliable and efficient storage is only a very small piece of the data lifecycle
  • Differentiation lies in having the most data, analysing it the fastest, discovering new insights, delivering new products and experiences not possible without it
  • Hyper-converged vendors are focusing on reducing cost by making storage just an integrated commodity function versus a platform for strategic advantage
  • It is time to banish SCSI for good
  • Legacy protocols most notably SCSI remain the archaic way we still speak to flash within storage arrays and to storage arrays over the network
  • Legacy storage vendors are retrofitting flash to decades-old systems, perpetuating the status quo versus embracing the true potential of flash
  • Nearly a decade ago the concept of converged infrastructure was born to gain advantages of an end-to-end tested stack
  • Our founding vision was to eradicate the spinning disk from the datacentre, but it turns out remnants of disk remain in storage architecture even after the disk is gone
  • Public cloud vendors are making it cheap to store data but expensive to use it
  • Pure Storage believes there is a place in the market for a pure-play data innovator, wholly-focused on driving data advantage
  • Storage does not exist as an island, it is deployed with compute, networking, virtualisation in the service of applications and users
  • The IT landscape today is dominated by vendors who are treating data as a compromise
  • The move to Non-Volatile Memory Express will be a multi-year transition, and we are committed to taking our customers there in a non-disruptive, investment-protected fashion
  • Vendors are all pivoting to low cost and low differentiation data strategies, versus investing in a platform to drive data advantage
  • We aim to deliver a data platform that enables our customers to build a new class of applications and to extract new insights from data – we call this data advantage
  • We are building a set of Non-Volatile Memory Express-enabled products that will allow mainstream Non-Volatile Memory Express adoption across a wide set of use cases
  • We believe Non-Volatile Memory Express should be an Evergreen evolution
  • We believe that Non-Volatile Memory Express solutions should incorporate all the critical data management features that customers need and expect
  • While storage has improved, data management processes have remained unchanged for decades and drive proliferation of unnecessary data copies
  • Yet in most datacentres today, data is also the key bottleneck

Pure Storage has set for itself seven innovation milestones and progressed through three. Matt Kixmoeller describes the recently announced four additional innovation storage milestones.

Excerpted from Pure’s Innovation Plan, Delivering Data Advantage by Matt Kixmoeller is Vice President Products at Pure Storage.


2016 Awards Banner

Latest Whitepapers

Whitepapers Covers
Whitepapers Covers
More Whitepapers
GITEX Advert