When is cloud genuinely a cloud?

I’m often asked, usually by older business executives with, at best, a rudimentary appreciation of information technology what all this cloud stuff is about. Often, they are being urged by younger, tech savvy executives and IT staff, to approve a “cloud first” or “cloud only” IT strategy. The problem is that the whole debate on moving to the cloud has become tribal, and one tribe, the digital natives, is zealously attempting to define what is “genuine” cloud and dismissing anything it sees from other tribes as cloud washing.

Somewhere over time, the term “cloud” has come to encompass both a consumption model and a technology. The consumption model refers to the pay-per-use, on-demand nature of accessing and utilising cloud services, e.g., Infrastructure-as-a-Service (IaaS), Platform-as-a-Service (PaaS) and Software-as-a-Service (SaaS). On the other hand, there is a technology stack that supports the infrastructure, virtualisation, automation, and distributed computing necessary to deliver those services.

So far so good and reasonably understandable. Now, let’s be clear, cloud does not equal Amazon Web Services (AWS), Microsoft Azure (Azure) Google Compute Platform (GCP). In theory, any service provider, including in-house IT department, can be a cloud provider. The US National Institute of Standards and Technology (NIST) has produced, in my opinion, the definitive description of what constitutes cloud computing. The document is short (3 pages), well laid out, and remarkably jargon free. On the basis of the NIST description you might indeed say that AWS, Azure and GCP equals cloud and that many other potential cloud providers would struggle to deliver genuine cloud services at a competitive price. Nowhere does the document preclude the use of particular technologies (for example, mainframes) in cloud services. Indeed, IBM now offer their Z Series mainframes running Linux in the IBM Cloud. But the cloud native tribe would be highly sceptical at best, and downright antipathic towards this as a genuine cloud technology.

It is in the underlying technology layers that enable cloud computing where the culture clash between tribes is most apparent, and most annoying. The antipathy identified above is mostly directed at vendors rather than specific technologies. Any cloud initiatives or product developments from companies like IBM, Oracle or VMWare are treated with the utmost suspicion by the cloud native tribe, despite the fact that many technologies used in cloud computing had their origins in legacy, often mainframe, developments. Message queuing technologies used in cloud computing most often trace their roots back to IBM MQ messaging in its mainframes. VMWare is synonymous with server virtualisation, that remains an important, if not essential component in cloud computing. Yet products, designed to be run in the cloud like VMWare Tanzu, which makes heavy use of Kubernetes, and IBM’s AIOps solution rely on a number of open-source tools.

I suspect that there is an element of cultural “not invented here syndrome” going on, alongside a suspicion that the legacy vendors, if I can call them that, do not really believe in cloud and are only trying to protect their existing user base. Yes, they are trying to protect their user base and realise that they have to believe in the cloud in order to survive. The story of IT has always been one about vendors trying to build and hold on to category and market power. Amazon, Microsoft and Google are no different. This clip from an Economist article about the impact of open-source Artificial Intelligence is quite instructive. “Now techies are abuzz about another memo, this time leaked from within Google, titled “We have no moat”. Its unknown author details the astonishing progress being made in artificial intelligence (ai)—and challenges some long-held assumptions about the balance of power in this fast-moving industry.”

At AWS RE:invent in Las Vegas in 2019, then CEO, Andy Jassy, said “The market segment that we address at AWS is the global IT market. If you look today, still the overwhelming majority of spend is on-premises.” It is no coincidence then that IBM, VMWare and others have made some of their services available on AWS, nor that, with products like AWS Outposts, it is putting hardware appliances directly into customer premises and making hybrid cloud more of a reality.

I realise I haven’t mentioned any other tribes other than the digital natives. The reality is that they probably don’t self-identify, but that is not to say that they don’t exist or that they are not prone to some over reaction and finger pointing of their own. A good example is repatriation. This has definitely become a thing in the last year or two, and I note a tone of triumphalism in some of the commentary that isn’t always supported by the detail on closer examination. There is not a general retreat from the public cloud, but some organisations are repatriating some systems back in-house. As often as not, the system in question shouldn’t have been migrated to the cloud in the first place (at least not in its existing form). It is incumbent on all parties to undertake due diligence and ensure there is a ‘know before you go’ mentality about the suitability of applications ability to perform effectively and efficiently in the cloud.

Yet, migration, refactoring, re-platforming for cloud, using all the same modern tools in use by the digital natives, are high-on their to do lists. If that is best achieved by working with vendors who have had experience and expertise in the Enterprise IT market for many years and includes hybrid-cloud, so much the better.

Cloud has undoubtedly been a huge agent of change in this century. Ongoing tribal disagreements are not going to fundamentally change the upward trajectory of cloud computing in all its forms. But those arguments are not going to help Enterprise IT to modernise and that won’t help any of the tribes.