20October2016

The Future of the Enterprise Cloud Operating System

VMware has announced the VMware Cloud on AWS. This new enterprise grade public cloud consists of the entire VMware vSphere/NSX/vSAN software stack and API’s along with the third party ecosystem for vSphere on AWS bare metal. Read the post to see how this new offering from VMware could reshape the entire cloud computing industry for the benefit of enterprise customers.

Cloud Operating System Strategy Changes Provoked by this Announcement

IF VMware makes a success of this new enterprise grade public cloud offering (note that all previous attempts at a viable public cloud offering by VMware have been at best a middling success) then the following may well occur:

  • AWS is now in the bare metal cloud as a service business. The other two vendors with significant on premise systems software businesses (Microsoft and Red Hat) may well want to do the same thing with AWS as VMware has done. This means Windows/Azure and RHEL/OpenStack on AWS bare metal as a service. So all major on-premise systems software platforms may end up running on AWS bare metal as a service.
  • Google and Microsoft may well decide to follow in the footsteps of AWS (as they have been doing for quite some time) and enter the bare metal as a service cloud business. If this happens there will then be three viable bare metal as a service cloud offerings and an entirely new market for enterprise grade cloud computing services may emerge.
  • Some of the cloud OS vendors might well choose to offer their software platforms on the Google and Microsoft bare metal clouds. VMware would be the most likely to do this as it would then force Amazon, Google and Microsoft to compete on price for hosting the VMware Cloud OS for VMware customers. But none of Microsoft, Amazon, and Red Hat are going to stand still and let VMware own the enterprise competent public cloud business.
  • If AWS concludes that being in the bare metal as a service business is a good thing with three customers, they may decide to open it up. This will have two very profound long term effects:
    • It will put tremendous pressure on the vendors of colo data centers. The pitch will be why own your own hardware in a colo data center when you can rent the hardware (and quite possibly better hardware) more cheaply and more flexibly from AWS and have that hardware provide you the certainty in performance and reliability that you require.
    • It may be the magic bullet that kills (or significantly dents) on premise enterprise computing. A lot of things still run on bare metal for very good reasons. A bare metal cloud could really eat into that installed base. Especially if AWS offers customers the ability to directly host things like their database servers on AWS bare metal alongside the hosting of their VMware Cloud instance on AWS. This would precisely mirror what enterprises have today. The only thing left on premise might be low latency and high frequency trading operations where distance over fiber to certain resources matters.
  • If all of the above comes to pass, then this sets up global thermo-nuclear war in the enterprise Cloud OS business. VMware, Microsoft and Red Hat will then have Cloud OS businesses that seamlessly span on-premise, hybrid and public bare metal cloud options. AWS and Google will be left with commodity (IaaS and PaaS) public cloud options only. This may well force them to get into the enterprise Cloud OS business themselves.
Enterprise Cloud Operating Systems – Long Term Winners and Losers

First of all it is likely to take 5 years for the above scenario to play out, and it may well play out very differently that laid out above. But if it does play out in any manner resembling the above, then the following winners and losers might emerge:

  • VMware has the best enterprise grade Cloud OS on the market today. VMware is the only vendor that has a robust trifecta of server virtualization (vSphere), network virtualization (NSX) and storage virtualization (vSAN). VMware is far ahead of its competitors in the crucial areas of enterprise availability, redundancy, control of physical resources, visibility into resource utilization, latency, and throughput, and the resultant ability to guarantee application performance. VMware also brings to the table an unparalleled ecosystem of partner vendors whose products collectively make it into the only enterprise grade Cloud OS on the market today.  This Cloud OS in combination with a choice of bare metal cloud execution environments will likely be a winner. The only major issue facing VMware is being owned by Dell, but if VMware is given the freedom to host its Cloud on OS on environments directly competitive to the Dell server business that may not matter.
  • Microsoft has a unique strength that no one else has. No other vendor has the largest on premise systems software business in the world (Windows) and a viable public cloud offering (Azure). If it enters the bare metal cloud as a service business then it will be the only vendor with a strong presence in all three segments. Microsoft also has the best position across the SMB and enterprise markets of the vendors. For Microsoft to be a real long term winner it must figure out how to combine its positions in on premise Windows with its position in the Azure cloud (offer Azure on premise for private/hybrid clouds), and also offer Azure on other vendors’ bare metal clouds should this make business sense. Microsoft must also create a viable enterprise grade Cloud OS out of a combination of Hyper-V and Azure in order to effectively compete with VMware in the enterprise.
  • Google has a unique strength and a major weakness. The unique strength is the ability to generate revenue through advertising which gives it options in its business model that no one else has. Consider free hosting of web applications in exchange for Google being allowed to place relevant ads in the side bar of the browser. Would you let Google put ads in the applications your employees use? In exchange for free hosting you might. The major weakness is how far Google would have to go to produce a credible on premise Enterprise Cloud OS and then to market and sell that OS. Google is probably the only company in the world with the software engineering talent required to transform its current commodity cloud offering into an enterprise grade Cloud OS. It remains to be seen whether Google has the vision and the desire to invest on the fronts of development, marketing and sales to pull this off.
  • AWS is the wild card here. With the largest IaaS/PaaS business, exceptional economies of scale, and a proven ability to produce high quality systems software, AWS is positioned to be a winner. But to be a long term enterprise winner AWS must figure out how to be more than a bare metal cloud as a service provider to enterprises with business critical and performance critical applications. AWS must make its Cloud OS into an effective competitor with the offerings from VMware and Microsoft to win in the long term in the enterprise on this front.  Just as is the case with Google, offering its own Cloud OS on its own bare metal cloud (no shared tenancy) along with operational visibility into the hardware would be a good start. But that alone is not enough. AWS must seriously invest in its Cloud OS in order to bring it to parity with the VMware Cloud OS as AWS lacks every single one of the enterprise grade features that makes the VMware Cloud OS stand out. AWS already has an enterprise sales force so it is well positioned to compete with VMware if it produces an enterprise grade Cloud Operating System.
  • Red Hat and the OpenStack open-source community will be challenged to make the required investments in product development, marketing and sales to win in the Enterprise Cloud Operating System market. OpenStack is therefore mostly likely to succeed only in organizations like service providers and the most sophisticated of enterprises that can find, hire and retain the people required to make it work at scale. If several large commercial entities rallied around OpenStack the story might be different but based upon activity in the most recent release, this is not the case.
  • XenServer is the odd man out. Citrix has neither the money, nor the will to make the investments required to propel XenServer into a leadership position in the enterprise Cloud OS market. The open source community has clearly rallied around OpenStack depriving XenServer of this stream of development talent. No one is stepping up to do the necessary marketing and selling. So it appears to be “game over” for XenServer.
OpsDataStore and the Future of Enterprise Cloud Operating Systems

Choices and competition are wonderful things for consumers and enterprises. The Cloud OS vendors are going to compete voraciously for your business and this will benefit enterprises worldwide. But the process of choosing what to run where will come down to:

  • Which option provides the required enterprise grade feature set?
  • Which option provides the required availability, reliability and redundancy?
  • Which option provides the required performance (response time and latency)?
  • Which option provides the required throughput (work done per unit of time)?
  • Which option provides the above most efficiently with the minimal amount of technical and human resources?
  • Which option provides the above at the lowest cost?
  • Which option provides the greatest flexibility and least amount of lock-in or switching costs?

All of the above are Data Driven Decisions. You need operating metrics from the entire stack of each of your applications to make these decisions. OpsDataStore and our ecosystem of partnering vendors is the only viable source of the required metrics organized by application system and business service.

Summary

With the VMware Cloud on AWS, and AWS offering a bare metal cloud service, VMware and AWS have provoked a strategic shift in the enterprise and public cloud businesses. Enterprises now face an expanded range of choices, and some very difficult decisions to make. AWS, Microsoft, Google, and Red Hat all face difficult decisions as to how to compete with what is for now the only enterprise grade public cloud offering.

For More Information Visit These Links: