Cloud’s next turning point: All-In or Over-the-Top

The next big Cloud technology movement isn’t a winner-takes-all battle!

Cloud technologies enable entirely new ways to compute, conduct business, deliver software, provide security, and more.

Trends can shake up entrenched vendors seemingly overnight. The biggest IT debates in recent history have pitted one technology against another similar technology. Startups and large vendors vie to win enterprise IT teams’ hearts and minds.  We debate the pros and cons of different technologies:

  •       VMware virtualization vs. Xen virtualization
  •       Public Cloud vs. Private Cloud
  •       Amazon AWS vs. Microsoft Azure
  •       Chef vs. Puppet
  •       Open Source vs. proprietary software
  •      Containers vs. virtual machines
Image credit: Wikimedia commons

Winner takes all battles

While Cloud has eradicated the need for other technologies, these technologies have not directly eradicated one another the way VHS swallowed up Betamax.  The shift has instead had an impact on everything above the data center level: how we deliver software, how devices connect, and even how we pay for IT services. Regardless of the type of cloud computing – public/private or SaaS/IaaS/PaaS – cloud is now the defining “design center”.

The New Dividing Line: “All-In” vs. “Over-the-Top”
The new split we are seeing is customers making a choice of going all-in on a specific cloud provider’s PaaS platform, or choosing to retain more direct control of their infrastructure by doing more of the infrastructure work themselves “over-the-top” of one or more cloud service provider’s IaaS platform.

We believe these choices are not being taken recklessly.  Those choosing “All-In” know they will have barriers to future migrations.  Those choosing “Over-the-Top” accept the fact that they will be managing greater complexity.  Cloud buyers and designers are making these decisions driven by the nature of the businesses they are building, choices which are determining the fundamental architecture of their business systems.

All-In:  Life in well maintained walled gardens
Cloud computing providers have created a massive cloud services portfolio. Look at the plethora of platform services available from Amazon AWS, Microsoft Azure, Google, and even “telcos turned clouds” offering up managed cloud services.

The AWS dashboard has services from data analytics to virtual desktops, and SSL certificate management to super-computing-sized databases. These types of services all aim to make cloud computing easier and more accessible for organizations.

With so many services and so many ways to interconnect them, why would a customer ever leave that platform?  Would you really expect your staff to build a hybrid PaaS-based application on BOTH AWS and Azure services?   Most likely (and reflective of what Cohesive sees in the market) customers are making a specific choice to use the PaaS platform of a single provider.  Thus, cloud-provider service platforms are becoming one of the most well manicured, expansive walled gardens in technology.

All-in:  Locked-In; Cloisters, not prisons
Organizations are going “All-In” and building apps and resources on a specific cloud PaaS platform.  In doing so, aren’t they getting “locked-in”?  Aren’t they getting more entrenched in that platform than even at the height of the Microsoft monopoly frenzy?   With the size and complexity of cloud-based applications running global scale businesses, why aren’t they consciously avoiding lock-in?

This generation’s IT buyers and business are not getting unconsciously tricked into a single service platform. While making a deep commitment to a single provider’s platform make migration of the business applications near impossible (at least today), these businesses are solving for a different outcome.  Worries about vendor “lock-in” are easily trumped by all the other concerns they are balancing: scale, reliability, time to market, minimizing code development, access to skills, etc., all dramatically minimized by their platform decision.

Current all-in cloud customers are commonly start-ups. These nimble young companies willingly get “locked in” to a provider, like AWS, in exchange for the huge cost savings and rapid time to market compared to building their own services from scratch.

All-in/locked-in is a choice; a conscious choice.  Vendor “lock-in” sounds like a huge negative, implying that organizations that get locked-in are making a critical mistake. For cloud customers, going all-in is more like taking holy orders or joining the military – it is a structured, limiting choice but one that comes with a network of support, security, and shared practices.

Image credit: wikimedia commons

Image credit: Flickr









 Over-the-Top: Break out of the walled garden
Over-the-top customers are using infrastructure as a service (IaaS) to deploy their own virtual infrastructure, then on top of that virtual infrastructure they deploy application infrastructure (middleware) and applications.

The over-the-top movement has already begun: Docker and micro services are proof that there is enterprise demand for re-drawing the lines and providing “computing power to the people”. Docker’s explosive popularity can in part be seen as customers wanting greater control of their environments, with containers representing “over-the-top computing”.  Increasingly they are also using over-the-top provisioning, automation and monitoring for these containers, separate and distinct from the provider’s IaaS offerings.

Prior to the emergence of containers, virtual networks, and virtual storage deployed on top of the public cloud, customers could only use the bulk compute, bulk storage and bulk transport offerings of providers like AWS or Azure delivered as IaaS.  Customers used this virtual infrastructure to deploy virtual machines with their applications and middleware, using block or object storage in cloud provider subnets.

Now over-the-top cloud customers still deploy virtual machines as a first step, but those virtual machines are used to build application-specific compute, application-specific networks, and application-specific storage, all controlled by service provider independent provisioning engines, portals, and monitoring; effectively their OWN virtual infrastructure – on top of one or more cloud provider’s virtual infrastructures.  They can then launch, own, manage and operate their business applications fully under their own control.

Users are taking back the power to deploy their own virtual infrastructure, followed by application infrastructure and their own application designs using the scalable, distributed, and powerful public cloud providers as a generalized resource.


Image credit: wikimedia commons

Over the top!

Over-the-Top: Attestable Control and Points of Presence, that’s why!
There are two very strong, obvious reasons we see customers going over-the-top today; one is for the simple need of points of presence, and the other is attestable control. 

Points of presence have been neglected, not fully recognized in the public discourse as a reason to choose cloud service provider A over cloud service provider B, yet it is quite common.  Businesses have many reasons for wanting to locate features, functions, and data in a particular geography, and if the current Safe Harbor brouhaha is any indication, this is going to increase not decrease.  Sometimes you need to deliver your cloud-based system in a particular location merely because “that is what the customer wants”.

Going over the top is less about avoiding lock-in, and more about demonstrating “attestable control.”  Customers in regulated industries (increasingly all industries), need to be able to explain how they control the semantics, workflow and dataflow of their business infrastructure.  For example the NIST Cybersecurity framework points to about 400 elements related to “network integrity”, almost ½ of all the items covered in the framework.   IT executives and Corporate executives are increasingly being asked to attest to this level of control in signed, legal documents.  When held to this level of accountability it appears to be quite important that the underlying public cloud IaaS (bulk compute, storage, and transport) does not have any meaningful interaction with, visibility of, nor control of, their applications.

All-In and Over-the-Top: Emerging Now
Given this context, it is not surprising that there are both emerging “All-In” and “Over-the-top” stacks.   While Cohesive sees customers that represent both ends of this spectrum, today much of the market is still in the middle.  That said, the trend we are seeing has us believe that this split in the market will become the defining IT and business decision made when creating and deploying business information infrastructure.

In follow up posts we will discuss in more detail the emerging stacks, as well as use-cases for managing and integrating to both All-In and over the top cloud architectures.


By: Patrick Kerpan