Geekonomics? Yes, we are paying homage to Freakonomics – the National Public Radio podcast series that reveals the hidden side of everything, applying economics theory and utilizing often hidden data to reveal the underlying truth. In Geekonomics we apply economic theory to all things IT to see what is really happening. Let’s do the math.

So What’s Happening?

Well, in the past week, we’ve seen several developments:

So What Gives?

Let’s analyze a slew of hypotheses about the market that are swirling around amidst the pundits and IT news sites right now.

Hypothesis: Did the public cloud take over from on-premises storage?

Premise: The major public cloud vendors are huge and they are growing, so they must be crimping the style of on-premises storage.

Our Take: Apparently not. Capacity shipments to the enterprise grew 5.2% overall, while the revenue dropped eight tenths of one percent. Yet IDC also reports that 80% of enterprises are experiencing some form of public cloud remorse and are actively repatriating 50% of their applications back to on-premises data centers.

Footnote: Very interesting to see that as they do, Enterprise Strategy Group notes that they disproportionately moving those applications onto a software-defined storage architecture.

Interim Conclusion: Public cloud is not crowding out on-premises storage, but it is accelerating the transition from traditional to modern storage architectures.

Hypothesis: Storage simply got cheaper.

Premise: We read every day about the volume of flash being manufactured, so perhaps economies of scale or the utilization of capabilities enhanced by flash – like compression and deduplication – are driving out costs.

Our Take: Not exactly. There are always ebbs and flows of different system configurations that impact overall revenue. In this past quarter, IDC noted that sales of all-flash systems – which have been more expensive – fell while cheaper hybrid systems trended up. Capacity up and revenue down seems to fit.

Hypothesis: HCI is disrupting dedicated storage systems.

Premise: The all-in-one approach is better, faster, cheaper, leaving traditional dedicated storage in the dust.

Our Take: Certainly not yet, and potentially never. There is no doubt that HCI is growing and has cultivated a solid following for more predictable use cases. And the public fights between Nutanix and VMdell (VMware and Dell) for market leadership, but the hard reality is that HCI remains a $10B market with a mid-double digit growth rate, a small fraction of a total on-premises infrastructure spend of $160B. At just 6% of the overall market, HCI can’t generate the massive swing in the overall market that we saw this quarter.

Hypothesis: It’s the economy, stupid.

Premise: Trade wars, the threat of global recession, and POTUS are at play, worrying those in charge of spend which caused a hesitation in spend.

Our Take: Read the Wall Street Journal and make that judgment for yourself, but our view is that blaming the economy is just a handy and simplistic means of avoiding the truth.

Hypothesis: Server-based infrastructure is finally taking over the market.

Premise: The industry and this blog have been talking about the merits of software-defined running server-based infrastructure for a long time, and the stated cost savings of moving to this approach are finally having an impact on the overall market spend.

Our Take: Our own analysis of the IDC numbers over the past three years indicates that server-based infrastructure has seen 4-5X growth compared to array-based architectures. We cannot yet project the overall financial impact on the on-premises market, but it dominates the public cloud markets since AWS, Azure, and GCP all run on server and not traditional arrays, and forward thinking enterprises are moving in the same direction. IDC sees “server SANs” growing faster than external arrays while William Blair’s Ader indicates that server-based storage will grow at 7X the rate of external arrays, and that’s exact the trend that Datera not only capitalizing on, but driving with a broader ecosystem.

So Where Does That Leave Us?

Final Observation: Just last year, Chris Mellor from Blocks and Files once called Datera the minnow taking on pretty much every other primary storage supplier. Chris was right – we are taking the traditional storage world head on with a new approach. While we may be a minnow, we are assembling a school of like-minded creatures to make it happen, including fellow minnows as well as sharks and whales like:

  • Mellanox (NASD: MLNX), the industry’s leading ethernet networking systems provider, set to become part of Nvidia (NASD: NVDA) later this year.
  • WekaIO, the world’s fastest file system software provider that is winning new high-performance file-based workloads and setting the pace in artificial intelligence applications.
  • Intel Corporation (NYSE: INTC), the technology force behind server-based infrastructure and storage class memory (SCM).
  • Scality, a leader in Gartner’s Magic Quadrant for object and file storage for many years running.
  • Cumulus, a driving force in networking software for the open systems that form the packet processing backbone of the modern data center.
  • Hewlett Packard Enterprise (NASD: HPE), the market leader in enterprise servers that are the building block of software-defined architectures and a key player in the enterprise storage and services markets that is “lifting” Datera into markets around the globe.

This culminated in the industry’s first software-defined leadership forum that Datera spearheaded in June.

So What’s Next?

Our next segment of Geekonomics will focus on the cost of public cloud versus private cloud and the cost of traditional IT versus cloud IT, which has been the subject of much debate.