There’s no getting away from the fact that cloud computing is having a profound effect on enterprise IT. This is due, in part, to once cautious executives now seeing it as a quick and easy shortcut to digital transformation. However, while some may find it a good fit, for others the brave new world of cloud computing is far from a perfect solution. Indeed many of the executives I talk to see themselves stuck in what Gartner calls the “trough of disillusionment” having rushed into the technology a little too quickly and without a full understanding of what they were buying.
It’s the cloud, stupid
There are, of course, lots of good reasons for moving at least some workloads to the cloud with by far the biggest lure being the simple on-demand delivery model on which most services are based.
Before, the cloud on-premise infrastructure teams would, typically, have to spend months justifying and purchasing the hardware, software and services they required. A task made even more arduous by having to second guess demand years in advance, with many over ordering ‘just in case’ for fear of having to go through the process all over again if they got their sums wrong.
With the cloud, that all goes out of the window. Need a couple of virtual servers, a database instance, or a hosted email server – next week, tomorrow, now? In the cloud, they can be up and running in minutes, along with all the storage and network connectivity needed to go with them and with no capital outlay required. Moreover, with a lot of cloud platforms, you only have to pay for the resources you actually consume, making the cloud – in theory, at least – the ideal fractional computing platform on which to host digital business workloads.
Unfortunately, it often is ‘in theory’ as in many ways, the public cloud is a little too quick and easy of a solution, and careful planning is needed if it is to deliver the fractional computing benefits many expect of it. Moreover, if you get it wrong it can be a lot harder to put right than might be imagined.
Choices, choices, choices
One of the biggest problems with the cloud is the sheer number and variety of platforms and services on offer. Last time I checked, AWS alone had over 150 products grouped under 20 different categories, from basics such as compute, storage, database and developer platforms to leading edge business analytics and AI tools. Add to that similar portfolios from other providers, plus the ease with which public cloud products can be purchased, and you have the perfect recipe for creating a mixed bag of platforms, services and apps which, while effective and perfectly manageable in their own right, don’t make for good bedfellows.
Low buy-in costs also complicate matters, making it easy for line of business managers to acquire public cloud products with little or no oversight from corporate IT teams (so-called Shadow IT). Which wouldn’t be so bad except that, unlike on-premise infrastructure where it’s possible to look inside the data centre to see what’s going on, management visibility in the cloud can vary considerably between vendors and across products, making it hard to even know what you’ve got, let alone stay in control.
New issues for old
Don’t get me wrong, I’m all in favour of cloud computing, but where once the cloud was seen as a way of fixing the issues around on-premise IT, I’m increasingly hearing of new problems being laid at its door. One of the most common is that while the cloud can, and does provide a cost effective platform for the development of new applications, when those apps are put into production the costs can skyrocket.
Beyond that, the executives that I have talked to complain that they find it much harder to keep track of costs when applications move to the cloud compared to on-premise deployments. In fact, many will secretly admit to having little more than a vague idea of the total spend, plus no understanding as to whether that’s ‘normal’, or what they can do about it.
Another common concern is the extra layer of complexity added by a technology widely promoted as a way of simplifying IT. Rather than making IT easier, companies often end up having to build additional specialist support teams, in many cases for each individual cloud platform, over and above those required for on premise infrastructure.
And lastly, there’s the little matter of cloud sprawl with isolated silos of data even more commonplace and much harder to eliminate where the cloud is involved.
We can fix it
Unfortunately, these issues aren’t something that cloud vendors themselves are likely to address, at least not in the short term. Hence why so many organisations are abandoning thoughts of cloud-only IT and taking a hybrid approach spanning on-premise as well as cloud platforms – just to get back some degree of control. At the same time, however, they also want the on-premise part to be as quick and easy to scale as the cloud – which is why, while traditional server and SAN sales are forecast to fall over the next few years, interest in hyperconverged platforms is surging ahead alongside growth in public cloud services.
It’s also why there’s been a flurry of activity around tools and technologies to not only empower C-level execs to keep a lid on costs but allow IT teams to move applications around and balance workloads across platforms regardless of the technology, vendor or implementation involved. These orchestration tools are the missing link in the equation to fix enterprise IT by enabling the enterprise to take full advantage of what the cloud has to offer on their own terms.