Rumours of the public cloud’s cost-efficiency are greatly exaggerated. The title of this post might sound hyperbolic, but many IT pros will have heard stories of their counterparts being marched out the door after maxing out their budgets with public-cloud charges. And when you compare the real, long-term costs of public cloud to private cloud, it becomes pretty obvious why.
This isn’t to say the public cloud is bad. In fact, it’s essential for any business that wants to innovate and grow quickly. Nor is this an excuse for IT departments to cling to their brownfields on-premise infrastructure. But wise IT leaders will draw a line in the sand and adopt a blended technology model in which public and private cloud stick to what they do best.
Is public or private cloud cheaper?
Businesses typically end up in the public cloud because they need IT resources cheap and fast. Most often, this is driven by line-of-business users who need to power a particular project or campaign faster than traditional IT can provide for. After all, a simple credit-card activation is far more palatable than going through reams of process, budget approvals, and waiting times for IT to actually provision what you need.
The problem with this is that public clouds lack the rigour that enterprises need. When I say rigour, I mean it in relation to aspects like backup, data recovery, security and compliance – all those things which are essential to not just good corporate governance, but also long-term growth. As a public cloud footprint grows, so do these costs.
The costs of public cloud grow in a strictly linear fashion as you add more and more IT resources. Private cloud costs, on the other hand, start out high because of your initial setup expenditure – a mix of both CapEx and OpEx because you’re setting up not just hardware and software, but also the team structures and processes to govern it. How do these compare over time? I’ve pulled together this graph based on the rough average costs across several public cloud providers per new VM, compared to the same costs per new VM on an EMC hybrid cloud:
You can see that with hybrid cloud, the initial costs per VM are steep, but they also steeply decline as the business scales up its IT footprint. That’s because the upfront costs – of new backup processes, automation procedures, security coverage for production data, and so on – start generating economies of scale the more units you add on. In other words, OpEx actually goes down for private and hybrid clouds, with the savings more than offsetting the initial CapEx costs that we’re often told are exorbitant and not worth paying.
Readers with a mathematical bent will realise quickly that if you graph TCO instead of cost per unit, public cloud actually leads to exponential growth in costs the more you scale:
The greater your public cloud footprint, the more risk you’re taking – and the less likely you’ll be able to extricate yourself. If you doubt that, ask yourself: do you know the costs of re-platforming from one public cloud to another infrastructure? IT managers all know the value of an exit strategy from any technology implementation, but we often seem to overlook this when it comes to public cloud – likely because adoption usually happens on a very time-sensitive basis.
So what happens when all of your business’ production data is in the public cloud, your costs have ballooned, and you can’t shift providers without massive disruption to business operating models? This is when the lack of an exit strategy often leads to the person responsible making an exit.
So what’s the solution?
The solution is NOT traditional on-premise infrastructure. Trying to transform brownfields IT is risky because most businesses don’t have visibility into the current costs of operation: even the big banks I’ve talked to often struggle to nail down their actual OpEx spend. In addition, on-premise infrastructure – and that includes private cloud – can never match the agility of public cloud. The solution? IT leaders should draw a line in the sand and define a new platform that gets the best of both worlds.
Hybrid cloud makes sense because it gives you access to automated, orchestrated, out-of-the-box cloud at a lower cost than a private cloud deployment. But at the same time, it also allows you to scale efficiently at the speed that the business demands. Public cloud maintains its place as the obvious choice for application and service development, but private cloud gives IT the resilience and rigour to take these apps and services to scale when they’re proven. One rule of thumb is that for anything under 100 units, use public cloud; when you start to hit 1000-plus units, you should already be running on a private cloud environment.
At VCE, I’ve seen many customers – banks, retailers, education bodies – succeed by adopting hybrid cloud models that explicitly compete against those offered by IT outsourcers. In a lot of cases, they actually perform better! A hybrid approach, when supported by converged infrastructure and operations, lets you speed up roll-outs with public cloud but support long-term growth with private cloud’s well-defined cost model. If IT powers your business – and for the majority of technology professionals, it does – you’ll want to trust your infrastructural and career performance to more than the swipe of a credit card.Tags: aws, azure, capex, cio, cmo, converged infrastructure, emc, employee, hybrid cloud, it costs, marketing, on-prem, opex, private cloud, public cloud, scalability, shadow IT, tco, vblock, vce