Why We’re Winning With All Flash, yet Not in the Gartner Magic Quadrant for Solid-State Arrays
by Gavin Cohen – Head of Product and Solutions Marketing


We announced our All Flash array on Feb. 23rd and it began shipping on March 16th. In just one and a half quarters, we have achieved unprecedented success:

  • Total of 188 AFA customers – more than any other AFA has achieved in this time
  • 55% of these are net new customers to Nimble, where the customer had no history or loyalty to Nimble, and bought purely on the product’s merits
  • Six of the AFA customer base are Global 1000 organizations
  • All the wins were highly competitive against EMC, Pure, HP, NetApp and others

Why is our AFA winning?

There are three reasons that customers choose our AFA:

1 – Our AFA is a generation beyond other AFAs

In building our AFA, we didn’t just take our Adaptive Flash array (hybrid) and replace the spinning disks with flash drives (though our competitors would love for you to believe this). Instead, we went back to first principles and spent more than two years architecting a system that could take advantage of the latest and future flash innovations, while simultaneously addressing the limitations present in other AFAs. Contrast this with other AFAs built 6+ years ago, an eternity in the world of All Flash. What worked six years ago, when SSDs were ten times smaller and ten times more expensive, doesn’t work quite as well today. In short, we wanted to build an AFA that was a generation beyond anything else available, and that’s exactly what we did:

  • Performance and scalability – By delivering both scale-up and scale-out, we can scale-up a single system to over 2PB at 300,000 IOPS and scale-out in a cluster to over 8PB at 1.2M IOPS. This is 5X the capacity of the biggest EMC XtremIO and 18 times that of Pure.
  • Designed for 3D TLC NAND – NimbleOS achieves a seven year lifespan on the latest 3D TLC NAND through intelligent and highly optimized flash endurance management and garbage collection software. But best of all, we also achieved 20% more usable capacity per TB of raw flash than other vendors who use the same SSDs (e.g. Pure).
  • Unconstrained by memory – AFAs require large amounts of memory (DRAM) to store dedupe fingerprints, block meta-data, snapshot meta-data, etc. in order to achieve high-performance. The amount of memory in the array directly impacts its performance and scalability. Nimble has built the industry’s most efficient AFA memory architecture. It requires 31 times less memory than XtremIO and 9 times less memory than Pure for a given amount of SSD capacity.
  • Better dedupe and compression – We use inline variable block deduplication and compression and, unlike competitors, don’t turn it off in order to keep up with heavy write workloads.
  • Higher resiliency – Triple+ Parity RAID allows our AFA to withstand three simultaneous SSD failures in each RAID group. In addition, the “+” refers to intra-drive parity which allows every SSD to independently recover from partial or sector failure. In contrast, other AFAs use the equivalent of Double or Double+ parity.

The following picture graphically shows how our architectural advantage plays out at 2PB of effective capacity. In order to match the capacity of a single Nimble AFA, competitors require massively larger configurations.


2 – InfoSight Predictive Analytics

Nimble is recognized as the pioneer and leader in predictive analytics for storage. InfoSight has been collecting and analyzing tens of millions of sensor data points per array, per day, since the first system shipped in 2010. The sensors collected by InfoSight span the entire infrastructure stack – all the way from the storage to the VMs. InfoSight provides three dramatic benefits to our customers:

  • Proactively informs and guides without alarm fatigue – Simply put, Nimble arrays just keep working optimally without the customer having to do anything. That’s because InfoSight is always monitoring the array with the goal of predicting and preventing issues before they impact the customer. Nine out of ten issues are discovered first by InfoSight. And, surprisingly, 54% of issues that Infosight addresses are non-storage issues.
  • Predicts future needs and simplifies planning – Predictive analytics and machine learning allow Nimble to accurately predict future needs, hotspots, areas of improvement, etc. In addition, InfoSight has a scenario planner that leverages installed-base learnings to model the requirements of anticipated future workloads.
  • Delivers transformed support experience from level 3 experts – InfoSight automates the tasks handled by traditional level 1 and 2 support staff. This has allowed Nimble to build a support organization made up entirely of level 3 experts who, on average, answer calls in less than a minute. Even the most complex issues are resolved rapidly because InfoSight has already collected all of the information required to solve the problem to root cause. Our storage industry leading Net Promoter Score (NPS) of 85 represents the strong confidence our customers have in how we support them.

3 – Unified Flash Fabric

And then there’s the magic of a single unified architecture that spans All Flash and Adaptive Flash (hybrid) storage arrays. The same version of NimbleOS runs on both platforms, giving users the same features, functions and management across All Flash and Adaptive Flash. More than 50% of Nimble AFA deployments were in unified configurations. This provides our customers with far more flexibility than an AFA-only architecture and serves two primary use cases:

  • Primary on All Flash with backup/DR/archival on Hybrid Flash – The economics here are compelling. Customers who want to put primary workloads on All Flash often find it cost prohibitive to use All Flash for secondary storage as well. However, with Nimble, primary data on an AFA can be replicated to a cost-optimized Adaptive Flash array at a third of the cost.
  • Single point of management spanning All Flash and Hybrid Flash – Some customers are not ready to put every primary workload on All Flash. With Nimble, you can cluster All Flash and Adaptive Flash arrays and manage them as one in a Unified Flash Fabric. You can then benefit from transparent data migration, where applications can be seamlessly moved between different array types. In this way, we can bring flash to every application – even the ones that don’t warrant the cost or need the performance of All Flash.

Why we’re not in the Gartner Magic Quadrant for Solid State Arrays

There are two key criteria used for inclusion in the MQ. The first is an AFA revenue achievement number which Nimble has already comfortably passed. The second is based on the general availability date of the product. Unfortunately, the GA date for the Nimble AFA was two weeks and two days later than the Gartner cutoff. Rushing the GA to meet the deadline was out of the question as it could have compromised reliability.

However, Nimble is a Leader in the most recent Gartner Magic Quadrant for General-Purpose Disk Arrays and is the only company founded in the last two decades to be in the Leader Quadrant. Many of the factors that resulted in Nimble being named a Leader apply equally to our AFA – our strong ability to execute and our strong vision. If you do want to know what Gartner thinks of our AFA, we encourage you to contact Gartner directly and speak with their storage analysts about Nimble.

In the meantime, we’ll continue winning with our AFA based on its own merits and look forward to being included in the Gartner Magic Quadrant for Solid State Arrays next year.