‘Cloud religion ‘ is the sudden and committed embrace of cloud computing. Like any tightly held belief system, cloud religion has its strict adherents and purists versus the more pragmatic mix and match practitioners, click to tweet. Like other belief systems, there are many pros and cons and philosophical questions: Do we create and manage our own private cloud of virtual machines? Do we rely on someone else to do this for us? Who will that be?

A brief personal history of cloud computing

My first exposure to what later became ‘the cloud’ was in early 2007. A company trying to recruit me had technology that could easily scale up or down on virtual machines scattered around the globe, provisioned with all the services they needed to run their own bespoke ad network, along with accompanying logging, monitoring and applications. They had DevOps guys floating around anywhere they could get reliable internet access 24/7. I was sold, and stayed with them for over seven years.

I’d been working with PeopleSoft and later Workday so I understood the issues around ‘virtual’ and the ‘no code’ mantra – one of the first aspects of cloud religion. Amazon had just determined that the scalable technology they’d developed to support their e-commerce could be sold as a service, and launched S3 and EC2  in August 2006.

It took years before competitors caught on to the cloud revolution. By then AWS was running the show, truly an  ‘internet OS ‘. This wasn’t without its consequence; when over half the internet relies on your services, when they go down, so does half the internet. I recall telling customers on multiple occasions, “it’s not us that’s down, it’s AWS…”.

So what’s ‘cloud religion’?

‘Cloud religion’ is the sudden and committed embrace of cloud computing, a realization of the need to convert from a classic internal IT server structure to the cloud. It’s often led by a technical evangelist within an organization, or an outside role who persuades an organization to convert. This often includes embracing a specific cloud service provider.

Like any tightly held belief system, cloud religion has its strict adherents and purists versus the more pragmatic mix and match practitioners, click to tweet. Like other belief systems, there are many pros and cons and philosophical questions: Do we create and manage our own private cloud of virtual machines? Do we rely on someone else to do this for us? Who will that be?

Pragmatists vs. purist

When organizations decide to move their data and other services to the cloud, they do so for many reasons, but often a company ends up with a mixed bag of solutions to address these problems, and that isn’t always a bad thing. Not everything needs to be supported with cloud storage or computing; that said, as cloud apps improve, it’s getting harder to justify a solution which isn’t mostly cloud. Who can resist the allure of allowing a third party to manage the overheads of data security and virtual machines deployment?. On this front, the purists have won!

This leaves the purist with just one question: which cloud platform to use. A few years ago, AWS was the only game in town. Google announced  ‘Google App Engine ‘ two years later, but it wasn’t complete until 2011, when it was released as Google Cloud Platform. Microsoft came into the game strong in 2017 with the Azure stack, and by leveraging existing enterprise adoption of its stack, had captured over 17% of the cloud computing market by 2019, while GCP lags at 6%. It’s worth mentioning Chinese powerhouse Alibaba, which is now China’s ‘internet OS’ with a market share of 5.6%, and old ‘Big Blue’ IBM, coming in at a meager 2%.

Sectarian conflicts 

Once you’ve got ‘cloud religion’, you still have many implementation options. You can have your own private cloud, using your own technology or an outside party, or any combination of these. You can just use the infrastructure services (IaaS), or various platforms or applications using PaaS/SaaS. And as far as outside parties go, there are the Big Three, and numerous other providers, which in total still control a third of the market share.

Home-grown versus third party provider

Yes, you can build and manage your own cloud. But should you? Third party providers now support such a wide variety of a la carte services, building your own bespoke cloud makes little sense. It’s difficult for most companies to build their own cloud technology stack and even harder to maintain it. It all depends on the size and skill of your company’s operations.

Once the adoption of cloud religion occurs and you’ve selected a strategy, the danger of simply selecting a provider and moving all your applications and other IT is to fall into the  ‘lift and shift ‘ strategy. Just sticking all your functionality on a cloud doesn’t mean you automatically gain all its benefits. Consider all the pros and cons before selecting a migration strategy. Doing a  ‘simple ‘ lift and shift can be far costlier than a more targeted approach.

Private, public or both?

Beyond just building your own cloud, you could implement a third party’s cloud computing services behind your own firewall, so you’re essentially servicing and managing your own Virtual Private Cloud (VPC). Established DevOps and IT may be tempted to set up their own cloud of virtual machines. Security is the main issue here, but that can be misleading -a VPC isn’t a magic security bullet. There are multiple drawbacks to adhering to only a single cloud provider, one of which is that you’ll end up losing the ease of integration of many SaaS services – internal SecOps can end up creating a real barrier to adoption.

There can only be one…

Many organizations adopt a single cloud provider, typically Amazon, Microsoft or Google. The driving force behind this isn’t always pricing or level of service, but a visceral reaction or one that’s based on pre-existing relationships. On a more practical level, the organization resources only support a single cloud.

Adopting a strict policy on what cloud technologies are allowed, especially if you’ve a VPC, can have many consequences. Many SaaS apps run on just one platform, so choosing this strategy limits services. Additionally, not all features are available in all platforms, although both Microsoft and Google have done their best to map to their Amazon equivalents as much as possible.

Sticking with a single cloud provider has a decent list of benefits and there are risks to using multiple cloud services. However, the growing consensus is that adopting a strict adherence to a single cloud platform is too prescriptive, and that a multi-cloud approach is the correct path.

The agnostics

Because cloud providers are largely the same , there isn’t a pressing need to restrict yourself to one strategy. Many CIOs deliberately choose a multi-cloud approach to preserve flexibility and freedom of choice. Similarly, many SaaS providers are also adopting a multi-cloud approach.

Ultimately, cloud religion isn’t just to embrace cloud computing, but a tactic and adherence to either a ‘closed’ or ‘open’ approach, click to tweet.

What’s The Answer?

Understandably, the devil’s in the detail,  and the issues brought up don’t necessarily have clear answers.  Notably:  

  • Why would I select and stick with a specific cloud provider?
  • What are the impacts and issues of adoption of single vs. multi-cloud strategy?
  •  When your data is in the cloud, what is the impact on your data ecosystem when cloud religions clash?

We’ll dive more deeply into these questions in future articles.


Authored by John Kuo (Head of Product) at Harbr