On September 30, 2016, Nutanix enjoyed a successful IPO, seeing its stock price jump over 130% on its first day of trading. In this Insight, we examine what does this signal for infrastructure vendors, particularly those that are heavily hardware focused.
Nutanix, in its offering defined itself as an “enterprise cloud company.” There is significance in this designation. While the cloud has been attractive to vendors, venture capitalists and customers alike, building trust in the Cloud by enterprises has taken some time. While Neuralytix does not consider Nutanix a cloud vendor, per se, we do see its offering as an enabler of cloud capabilities. Its products can be used by service providers to simplify operations, reduce cost, and allow for scalable multi-tenancy.
Additionally, while it is most often delivered on prescribed hardware, Nutanix’s products are software defined datacenter software; and Neuralytix believes that the success of Nutanix’s IPO rested on the financial market’s perception of the software-defined market. Much like the dot-com days, when every vendor wanted to be a software company to garner a better valuation, today’s equivalent is to be a software-defined something company.
Other notable vendors in the space include SimpliVity and Dell EMC.
The interest in software-defined anything is not altogether unjustified. For the most part, hardware infrastructure is essentially commodity hardware. This is unquestioned when it comes to servers. But even storage and network are essentially disks and ports driven (typically) by x86 CPUs running some software (typically) on a Linux stack. Some proprietary hardware may be included for highly specialized operations, but, hardware delivered, or software-only, a good majority of infrastructure today is software-defined or software-driven.
Software-definition makes a lot of sense. It is easier to update or customize software over ASICs. Security updates and patches can be performed in near real-time. The density of cores and the performance of contemporary x86 CPUs are far superior to the bandwidth of Ethernet, hard disk drives, and even SSDs. Advanced computations and features such as deduplication or compression can be performed inline very limited impact to performance. For storage, basic functions such as RAID can now be performed in software rather than dedicated hardware for less cost and in many cases higher performance.
So where does this leave hardware infrastructure vendors?
Fear mongers will predict the death of hardware infrastructure products and vendors. But Neuralytix believes that this is highly unjustified. History has taught us time and time again, that enterprise customers are reluctant to change. Particularly where customers have integrated processes around a particular vendor’s hardware, it makes those products very sticky, and often takes up to three refresh cycles (roughly 10 years) to make any substantive change. This is a good reason why the EMC Symmetrix continues to enjoy the success it does, as well as mainframes and tape libraries.
That said, most of the hardware infrastructure vendors have already made commitments or wholesale changes to their designs. Switches and directors have already introduced many more software components over hardware changes, many of which are based on x86 processors. Storage vendors have also moved to a primarily x86-based design. Even the EMC Symmetrix is x86 based, and software driven!
Neuralytix does not fear that vendors will lose market shares just because they sell a “hardware” product. Switches and directors will continue to be sold, much as they are today. Storage will continue to be sold as a separate component for many years to come.
However, what we envisage over the next several years is a massive consolidation in the storage market. We believe that there are far too many vendors in the market today. Small product differentiation is insufficient for vendors to survive. We see this in both Violin Memory and Nimble Storage. Both vendors offer slightly different forms of distinctions. Violin Memory’s claim to fame is being a first-mover advantage, and the fact that they used proprietary solid state memory chips, as opposed to almost all other vendors that leveraged finished SSDs instead. Nimble Storage, as Neuralytix has criticized, offered a “me too” product – it offered every feature imaginable, at a slightly lower price, and slightly higher performance. In other words, it leapfrogged the market at every iteration, but offered no organic innovation. In Neuralytix’s opinion, neither company will be able sustain market retention.
Who will be the winners?
So, which vendors will come out winners in this market? In Neuralytix’s opinions, those companies that can deliver a “datacenter in a box” will be the winners. The most significant play is witnessed through the Dell EMC merger. Enterprise customers want to focus on insight and innovation, not infrastructure. Acquiring infrastructure in standardized units makes perfect sense. But the standardized units cannot simply be a bundle of components. Any system integrator or reseller can bundle components for a sale. To truly meet the real demands of an enterprise, the standardized units must be curated and most importantly, integrated via a singular management platform – bringing us full circle to Nutanix.
Vendors of tightly integrated datacenter solutions, including Nutanix, are going to be the ultimate winners in the long run. Highly profitable winners will be the ones that also provide the necessary services to deploy, provision, operate, and maintain these systems, which brings us full circle back to the concept of Cloud.
So, in describing itself as an “enterprise cloud company,” Nutanix has hit all the necessary terms necessary for what a successful infrastructure vendor needs to be moving forward. Neuralytix also acknowledges Dell EMC (VCE) and SimpliVity, amongst others having this winning formula.
How does a vendor become a winner?
For most vendors, the first step is done – converting their solution to be a software-defined/software-driven product. The next step is more difficult. Partnering and developing a single management, orchestration, and provisioning platform that is in the mutual interest of all parties may not be easy. For those who wish to focus in on the VMware ecosystem, it is somewhat easier, given that so long as there are management plugins that work with vSphere, that’s a major step in a unified, singular management platform. But even then, how does a vendor distinguish itself above all the other qualified and certified solutions in the VMware ecosystem?
Even more difficult will be how to buck the trend started by hyperconverged infrastructure (HCI) vendors who are driving out disparate and distinct networking and storage hardware infrastructure, and integrating everything into a cluster of 2U/3U form factors? For example, does a storage company partner with an HCI vendor to come up with a solution similar to Pivot3’s that distinguishes storage from compute/network?
All these questions having been raised, it does not detract from the fact that the infrastructure market is due for a consolidation. It is our considered opinion that there are too many vendors, particularly in the storage space. That said, we believe that the market will find its own equilibrium over the next five years through 2020.
The implication that Nutanix’s successful IPO demonstrates, is that for infrastructure vendors, product marketing efforts must move towards software-defined something. Cloud-like deployment, provisioning and management are key features that customers will be looking for. Integration with adjacent technologies will also be critical for success. Ultimately, to sustain success, partnerships will be the essential to a vendor’s long term survival.