A hybrid cloud strategy has plenty of appeal, but “it’s free” has never been a part of it.
Rather, from a financial standpoint, the attraction is often tied to the pay-for-what-you-use model, which among other things enables organizations to shift IT spending away from an CAPEX model (where you buy hardware, software, and other needs up front, regardless of future usage) to an OPEX model, where you pay over time for the resources you actual consume. The latter helps minimize the risk of buying a bunch of infrastructure that you may only need intermittently, for example.
But moving from one form of spending to another doesn’t automatically mean you’re reducing your spending. In fact, without smart planning and management, you could find yourself in for an unpleasant surprise when the monthly bills arrive.
“One of the attractions of cloud computing, especially public clouds, is that it’s quick and easy to spin up additional resources if demand spikes or you need temporary capacity for some experiment,” says Gordon Haff, technology evangelist at Red Hat. “However, it’s also easy for costs to spin out of control. And this is doubly true in the case of complex multicloud or hybrid cloud environments.”
[ Learn the do's and don'ts of hybrid cloud: Get the free eBook, Hybrid Cloud Strategy for Dummies. ]
That doesn’t have to happen, though. Many surprise cloud bills are the result of not keeping close enough tabs on your environments and usage, for example. The “classic” example here is the dev or other IT pro who spins up a VM on a cloud platform and then forgets to turn it off after its purpose has been fulfilled. It’s the cloud equivalent to never turning your lights off at home, though potentially an order of magnitude more expensive.
[ Get a primer on hybrid cloud: How to explain hybrid cloud in plain English ]
Five hybrid cloud cost misunderstandings and how to avoid them
Unexpected costs can also happen as a result of a simple misunderstanding: “We didn’t know” being a common beginning to someone’s explanation for a surprise bill, or costs that seem to continuously rise. We’re here to debunk some of the biggest misconceptions to help you proactively manage your hybrid cloud and multicloud costs.
1. "My bill will explain everything."
You cloud bill(s) will tell you what you owe, but not necessarily why you owe it. Put another way: You need to be able to see and comprehend the big picture of your spending, not just a dollar amount.
“Costs need to be understood and monitored at a global level,” Haff says. “Cloud provider pricing is complex and the cost of services like data transfers may even help determine the right architecture for an application, where it is run, and where its data is stored.”
Just waiting for the bill to roll in each month may be indicative of a siloed approach to cloud management. That can lead to budget waste, says Manish Srivastava, VP and GM of IT asset management at ServiceNow.
“IT leaders need to look at their cloud assets holistically in order to effectively manage their cloud spend,” Srivastava says.
With a narrow or siloed approach, you can probably say “this is how much we spent last month (or last year.)” But you want to be able to explain who spent what, and why they spent it.
“By taking a holistic approach, leaders can identify the top consumers of cloud resources and make necessary adjustments for cost optimization,” Srivastava says. “After all, knowing who is spending money and wasting money allows for better accountability across the organization.”
[ Working on hybrid cloud strategy? Get the four-step hybrid cloud strategy checklist. ]
2. "Cloud is always cheaper."
A hybrid cloud architecture doesn’t automatically match the right workload to the right environment from a technical perspective; the same holds true from a financial perspective.
“‘Everything is always cheaper in the cloud’ – not true,” says Lenley Hensarling, chief strategy officer at Aerospike. “Everything is more flexible, more elastic [in the cloud] – that is true. But if you have stable workloads it is often significantly cheaper to run these in a co-lo private cloud or on-prem.”
As Haff mentioned above, data transfer costs are a prime example of a potentially overlooked cost – one that can be mitigated with proper planning and design, but not ignored. Egress costs – meaning the charges associated with moving data out of a particular cloud environment – are a big one to watch.
“Egress costs happen [when moving] from cloud to on-prem, but there are even charges when moving data between zones and regions,” Hensarling says. “Careful design can minimize these costs.”
You might also need to rethink how you’re measuring costs and ROI over time; worshipping too faithfully at the OPEX altar might lead you to ignore the reality that, over time, your pay-as-you-go spending will catch up to what you would have spent in your own datacenter.
It’s not so much “getting it wrong,” but something more fundamental, says Gokul Rajagopalan, director of products at Vectra: “Cloud costs are just a blind spot for most.”
It’s a matter of knowing what you’re using and why – rather than assuming everything gets cheaper by moving workloads to a cloud, just because you can. Similarly, measuring cloud ROI isn’t just a strictly quantitative matter of CAPEX versus OPEX, especially if you’re conveniently ignoring secondary costs, such as ongoing maintenance or upgrades in your own datacenter.
Moreover, your reasons for running something in a particular environment (including more qualitative measures, such as access to newer technologies) matter. Performance, compliance, security – all of these and more can be significant factors. Analyze your hybrid cloud strategy through a cost-benefit lens, rather than solely through a cost lens.
[ What are the biggest benefits of an open hybrid cloud strategy? Learn more about Red Hat's point of view. ]
3. "We know our platforms inside and out."
Even advanced teams can run into this one: Cloud IQ can get stale, because the landscape (and any given platform in that landscape) changes with regularity. On the one hand, this can be beneficial – you have access to new capabilities faster and more frequently than you might otherwise if you only went with homegrown solutions.
If you’re not keeping tabs, though, you may be missing out on newer options for doing things that may be more efficient and/or less costly than your current approach.
“Of course, unexpectedly large bills are commonplace. But another frequent gotcha I see is knowledge obsoletion, given the speed of innovation of cloud providers,” Rajagopalan says. “Every six months, there is a better-faster-cheaper way to do what you’ve been doing. One must continually reevaluate their use of the cloud to see if there are new cloud services or features that they should leverage. These optimizations can result in significant improvements in performance or reductions in costs that are easy to miss.”
4. "What's one more provider?"
Hybrid cloud and/or multicloud should not be cover stories for “unchecked acquisition of new cloud services,” though that is the reality (or part of the reality) in some organizations.
“Most large organizations use services from multiple providers,” Haff says. “There are good reasons for this including redundancy, unique capabilities, cost, data governance, acquisitions, and more. However, don’t unthinkingly proliferate complexity. Doing so makes it harder to predict costs and may also increase costs.”
[ Also read: Hybrid cloud strategy: 5 contrarian tips ]
An apparent advantage of cloud platforms and services – fast, flexible access to business-critical technologies that might have otherwise been out of reach or huge operational burdens – can lead to waste if no one is minding the store. Add platforms and providers because they solve a problem or fulfill a business need, not just because you can.
“A continuous effort needs to be made at every level of an organization using cloud IaaS/PaaS as well an ever-growing list of SaaS apps to focus on cost reduction and best practices,” Srivastava says.
5. "We only pay for what we use."
Cloud economics are often grounded in this principle, and it’s fundamentally true. And on the flipside, it’s not often feasible or sensible to make major data center investments just in case you might need some extra capacity down the line.
But this principle can also disguise one of the most common sources of overspending, especially in public cloud: When you are, in fact, paying for something that you’re not using. Again, the go-to example here is the zombie instance that never gets turned off, which means the meter is still running, too.
“It’s easy for developers to just fire up the biggest virtual machines and never delete data,” Haff notes.
Only paying for what you actually use is a wonderful principle; just make sure you’re actually using what you’re paying for. You need enforceable strategies for making sure people turn the lights off when they leave the room, so to speak, and in general get rid of stuff they no longer need.
“Establish default policies and shut down services that aren’t being actively used,” Haff says. “Chargebacks are one common approach to increasing cost transparency and accountability.”
[ Learn more about hybrid and multi-cloud workload strategy: Get the free eBook, Multi-Cloud Portability for Dummies. ]
Subscribe to our weekly newsletter.
Keep up with the latest advice and insights from CIOs and IT leaders.