A couple of weeks ago, I had an amazing and interesting day that got me thinking about what might be the future of Enterprise DC infrastructure.
The first amazing thing was watching our daughter’s reactions when my wife and I bought her first phone; she was so excited. The second interesting thing happened that evening as I watched a primetime NBA game. During a commercial break, for the first time, I saw a Dell Technologies ad for integrated, one-stop-shop Enterprise DC infrastructure.
I then started to wonder if the DC ecosystem would evolve as smartphones have. But before we predict the future, let’s briefly recap history.
Convergence Driven by the Smartphone
My kids laugh at me when I tell them what it was like before the smartphone. They just can’t believe that I used to carry an array of disintegrated, single-purpose devices including my flip phone, music player, pager, camera, portable navigation unit, PDA and my watch. Today I just grab my smartphone since everything is converged within and around it.
Today’s Smartphone Ecosystem
Today, the smartphone industry has consolidated into three ecosystems that represent two different business models.
Single-Source Solution – Apple: From the beginning, Apple pursued a walled-garden, fully integrated, closed-ecosystem approach. They provide the devices, the device OS and many if not all of the required infrastructure services and fully control the user experience. They have facilitated a large, value-added, app ecosystem on top of their stack, which continued between what is an app on top and what is an infraservice that they should provide organically.
Over time, they will no doubt subsume whatever they can into their infrastructure to maximize their ability to monetize services across their installed base (as they have done with SMS/IM, music purchases, navigation and so forth).
Integrated Systems – Google and Microsoft: The alternative model to single source is the integrated approach pursued by Google (and Microsoft even though Windows Phone is below 1% in market share). This model involves the delivery of the device OS, infrastructure services and software stack from one vendor integrated with devices that are provided by other vendors as well as organically provided by the stack vendor.
This results in Android and Google cloud services integrated onto their own Pixel devices as well as integrated with third-party devices from Samsung and others.
What happens next? Today’s ecosystem is metastable at best, with significant change likely in the near term. Apple has proven the success of the walled garden, garnering more than 90% of profit share with about 30% market share depending on whose numbers you believe.
As growth slows due to smartphone saturation, one has to assume that both Google and Microsoft migrate toward a walled garden where they eliminate their device partners to maximize profit and user experience. Microsoft has the existing mind share and channel relationships into Enterprise to do this with a narrower focus on business customers. Google now has competitive handsets, an emerging channel with phone companies and a higher profile brand to do this with a consumer focus.
This would be a devastating development for Samsung and others that rely on the Android stack. The rise and fall of Nokia in cell phones might be a history these companies are destined to repeat.
Enterprise Data Centers
We are in the middle of a convergence cycle in the data center—about where phones were a decade or so ago. We used to bring best-of-breed, discrete technologies from a myriad of different vendors together to build the Enterprise data center.
We would bring a compute vendor, several storage vendors, a storage networking vendor, an Ethernet switching vendor, a hypervisor vendor, an orchestration vendor, a backup vendor, several security vendors and others into the data center and use our IT staff and/or consulting services to integrate and operate it all.
This era is ending with the introduction and rapid growth of hyper converged infrastructure. We now can get storage, compute and networking integrated within one product.
Single Source Solution – Dell Technologies: The ad I saw on primetime was very straightforward. Dell was promising an integrated, cost-effective, full DC solution. For the moment, they still partner with others for some pieces of functionality, but it is clear that they aim to offer it all sooner rather than later, including security, backup, vSwitching and so forth.
The main anchor point they use with Enterprise customers as they rotate towards this walled garden is the VMware orchestration and virtualization stack since it is the stickiest piece of the solution. So this is the Apple equivalent.
Integrated (Software) Systems – Nutanix: Nutanix has led the charge on hyper converged and has done a great job in getting to IPO through integrating underneath the VMware stack.
But they clearly understand the existential threat of the walled garden mentioned above, and are implementing their own KVM-based software stack as fast as they possibly can to try and establish their own walled garden, albeit sitting on commodity hardware provided by a set of third-party partners. So this is the Google equivalent.
Integrated (Hardware) Systems: There are many traditionally hardware-oriented vendors selling into the data center. On the compute side, there is Lenovo, HP, Cisco, Huawei and others. On the storage side, there are folks like NetApp and Pure. On the networking side, you have folks like Cisco, Juniper, Brocade, Huawei and more.
They have all had partnerships to deliver an integrated stack on and including their hardware. While there are other stacks, such as OpenStack and System Center, 95% of the Enterprise DC revenue is still under the VMware stack. So again, the Dell walled garden is a major threat. These folks face the same risk as the non-Google Android handset manufacturers—that is a precipitous decline in market share and profit.
Microsoft: Microsoft, as with smartphones, is a unique entity with regard to DC infrastructure. They have tried with little success to challenge VMware in orchestration and virtualization sitting on third-party partner hardware. While they have some socket share, they have achieved little revenue share with Hyper-V.
Simultaneously, they have been very successful in building up Azure. Therefore, they are on their way to establishing a duopoly with AWS in public cloud (time will tell if Google can join this club). Their on-premise strategy now seems to center around delivering Azure Stack (again integrated on top of partner hardware) to Enterprise whereby all orchestration and management is delivered from the Azure cloud.
The continual delays in the introduction of Azure Stack are a testament to how difficult it is and how close integration is needed with the underlying platform(s). As with phones, I would imagine that Microsoft is considering delivering the complete stack (hardware included) to make it easier and to enhance profitability.
Summary: The next phase of this will be fascinating. Hardware-oriented Enterprise vendors are facing a perfect two-front war: public cloud on one side and Dell Technologies on the other. Despite the strong growth rates of global demand for compute and storage, the Enterprise DC market is fairly flat. 451 Research predicts a CAGR of 2% to 3% for Enterprise DC and a CAGR of 22% for public cloud over the next several years.
How will these vendors evolve and survive? There seems only to be two options. The first option is to cede the Enterprise DC market to Dell and refocus efforts to sell to hyperscale cloud customers, reorganizing and optimizing costs to live on the slim profit margins. The second option is to develop a next-generation, full-stack solution that can both compete with the Dell walled garden and can offer a full cloud experience to the Enterprise customer so that the solution is as consumable as public-cloud services.
This lack of cloud experience is probably the enduring Achilles heel of the Dell solution given the on-premise history and DNA of the VMware software stack.
It will be interesting to see who among the main Enterprise DC vendors will manage this transition and thrive, and who will fail to act and become irrelevant in the Enterprise DC—just as Nokia became irrelevant in mobile phones.
-Michael Beesley, founder and chief technology officer, Skyport Systems