Wednesday, November 1, 2017

A Tale of 3 Cloud Strategies - Part III

How the major vendors are vying for hybrid cloud dominance

Over the past few years it’s been an exciting show to see the Big 3 cloud providers jockey for strategic dominance, each with a different approach. As an IT professional, it’s useful to understand each of these in context as you try to determine which horse(s) to ride and what bets to make.

To recap how I’ve been tracking some of the evolution:
  • In Part I (back in 2013) we observed AWS experimenting with expansion by reaching down into the enterprise by licensing its APIs (remember Eucalyptus?) and building a VPC (which evolved into Direct Connect). Meanwhile VMware was experimenting with growth by partnering with independent cloud service providers, a precursor to their failed vCloud Air approach – an attempt to reach from the enterprise up into the cloud. 
  • In Part II (later in 2013) I added a 3rd player to the scrum, when Microsoft introduced the concept of the Azure Pack to their Server stack, a strategy to reach down from the Cloud into the enterprise and bridging the two with a common set of APIs.
In essence, AWS, VMware and Microsoft all looked at ways to expand their presence to create easy-to-adopt hybrid cloud strategies that would lower barriers-to-adoption... and hopefully accelerate enterprises landing workloads onto their stacks.

Enter VMware Cloud Foundation: 

Recently, VMware announced Cloud Foundation, a new strategy to (in my opinion) replace vCloud Air.  In concept, this approach is not unlike Microsoft’s Azure Stack strategy.  VMware is building a common cloud workload management platform that incorporates their SDDC that will operate with both public clouds and on-premises VMware implementations... thereby lowering the barrier-to-adoption of hybrid workloads based on VMware technology.

While Cloud Foundation is similar in concept to Microsoft’s Azure Stack, the two companies took very different approaches to implementation.  Microsoft in essence took the Azure API set, and embedded it into their on-premises server software.   In contrast, VMware took their (mostly) on-premises API sets and is embedding them to public cloud provider offerings.

VMware’s approach is interesting – and potentially very successful. First, they’re cutting deals with major public cloud providers like IBM, AWS, and Azure, so they are able to embed their virtualization stack on top of these public cloud platforms.  And next, they’ll (likely) begin to work with their ~ 4,000 cloud service provider partners to do the same, enabling a pervasive set of common APIs across thousands of providers large and small.  If you buy-into the VMware view of the world (and, some say, the “vTax”) this could give you the ultimate degree of common cloud choice... and put the reach of VMware across as many clouds and on-premises infrastructures as Azure.

Snapshot: Where we are today
As I see it, we now have the following competitive landscape and strategies: 

AWS: 
  • Strategy: Abandoned licensing APIs; Now 100% invested in public cloud presence, and in building-out a dominating set of services – to make their APIs into a de facto “cloud OS”
  • Play: Focus on the single best public cloud IaaS platform and PaaS services for developers and enterprise workloads
  • Business Model: Make money from services and from running workloads
Azure:
  • Strategy: Expand from Cloud  Enterprise. Take their public cloud APIs and duplicate them on-premises within their installed base within the enterprise.  Further helps blur the line between on-premises and cloud, as-has Office365.
  • Play: Reduce/remove barriers-to-adoption of the public cloud by first encouraging API adoption on-premises
  • Business Model: Make money both from server software (traditional) as well as from cloud workloads landed on Azure.   
VMware:
  • Strategy: Expand from enterprise  Cloud. Introduced Cloud Foundation, coupled with SDDC. Initially work to deploy on major public cloud providers, then expand platform to 1000’s of cloud service providers.
  • Play: Encourage pervasiveness of the platform by expanding its reach onto public clouds and into managing heterogeneous hypervisors and containers
  • Business Model: Make money from becoming the central management software (while partners make money on workloads)
Other Players? 

There are still some players out there that are still to be reckoned with.  Namely Google Compute Cloud, as well as IBM/SoftLayer.  While I’m not yet aware of any Google GCP approach/strategy, IBM has just announced their “IBM Cloud Private” approach, based on a number of open-source technologies. This also seems to be a hybrid container/deployment management model – and we’ll look to the future to see what traction it gains. 

In summary... 

As go most major industries and technologies, it appears that industry consolidation has pared-down the major cloud players to 3-5.   So, I’ll close by pointing you of my favorite, prescient Blog back from 2006 by Greg Papadopoulos, then Sun Microsystem’s CTO – Why the world only needs 5 computers.  True Dat.

No comments: