Cloud is all the rage in IT infrastructure—but it gets expensive. Fast.
If you don’t break out of legacy thinking, you can lose your shirt on cloud systems. That’s the case whether you’re managing those systems in house or outsourcing to a cloud managed services provider.
The key is to plan strategically and make the cloud work for you—not vice versa. You’ll want to optimize costs, both in the planning phases and on a regular basis, to get the most out of your cloud strategy.
In this article, we’ll cover common cloud hosting pitfalls. Then we’ll talk about cost optimization strategies.
Common financial pitfalls of cloud hosting
There’s no way around it: The cloud world requires a different mindset than on-premises hosting.
For organizations that don’t have deep experience with the cloud, it’s easy to fall into several common pitfalls. Here are the biggest mistakes to avoid.
1. Forklifting everything into the cloud
On-premises delivery is built on the concept of a physical server running 24/7. While you can replicate this model in the cloud, it’s rarely a good idea. “Always on” gets incredibly expensive in the cloud.
How do you avoid this?
Don’t forklift your existing architecture into the cloud. Rather, build your cloud strategy around services, not servers. More on that in a moment.
2. Not analyzing cloud options (and not planning strategically)
To succeed with the cloud, you want to take a step back. Rather than analyzing servers, you want to look at things in terms of essential functions. What functions do you need to fill? You’ll want to audit your needs in these terms, then analyze the available cloud resources on the market and find the right fit for each function. A cloud service, not a server, will usually be the best option to fulfill a given function.
3. Locking yourself into a cloud contract that isn’t right for you
If your cloud vendor offers you good pricing in a long-term contract—and if you’re buying services rather than servers where it makes sense—then it may be wise to lock up your rate in a 3-5 year contract.
Like a fixed-rate mortgage, this type of contract gives you clear visibility into your future unit cost for the cloud. But just like a fixed-rate mortgage, it can hurt your wallet if cloud unit pricing drops below your contract rate.
4. Choosing fluctuating cloud pricing… and not monitoring it
Some organizations don’t need the long-term cloud contract. Their cloud needs fluctuate, and so they would rather pay that market rate and manage their usage themselves.
That’s great—if you’re monitoring your spend periodically. But cloud cost optimization takes time and resources, and it really requires deep cloud expertise. Most midmarket companies don’t have that cloud expert on staff who really knows how to optimize costs regularly. That thought leadership is one of the biggest reasons to hire a managed service provider—but we’ll get to that.
5. Not understanding what a modern application looks like
In the on-premises world, applications come bundled with their databases. You can’t have one without the other. That database dictates operating system dependencies, and those dependencies dictate your hardware needs. The whole thing is monolithic.
Believe it or not, this model is still widely used—even by big-name software vendors. While they may have a cloud version of their legacy software, these vendors haven’t really built modern applications. Hosting legacy-style applications in the cloud gets very expensive, very fast.
Maybe you have an application that depends on a SQL database, and that database requires an operating system. What would it take to remove those dependencies and host SQL as a service in Azure? Then you’re paying for transactions rather than something that’s running 24/7. In most cases, it will cost pennies—or even fractions of a penny.
6. Not analyzing TOC (total cost of ownership) for cloud vs. on premises
This isn’t a simple question of, “How much does it cost in the cloud vs. buying hardware?” If you buy your own hardware, you’ll encounter hidden costs for upkeep, patches, and general maintenance. Cloud vendors cover these functions and bundle them into their fee structure.
How do you get the full picture here? You’ll want to examine both implementation costs and ongoing costs for each alternative, calculating total cost of ownership across the usable life of the system. Again, there are no right answers, only appropriate compromises.
Cloud cost optimization strategies
Luckily, you can optimize your cloud costs. All it takes is attention to detail, a solid plan, and a little cloud expertise. Here are the strategies that we implement most often for clients.
1. Think in terms of services, not servers
We’ve mentioned this a few times already, but we can unpack it further.
The first and best way to optimize cloud costs is to take a services approach. Cloud services run and accrue cost on an as-needed basis, making them far cheaper than cloud servers.
When you decouple applications from databases, operating systems, and hardware, a whole new world opens up. Cloud services can deliver the functions you need, when you need them, without the massive expense of always-on cloud hardware. For the vast majority of applications, this is the #1 way to optimize cloud costs.
2. Take a lean approach to disaster recovery resources (and model those costs ahead of time)
Your backup and disaster recovery plan is a great place to optimize cloud costs. Your company may not realize that you can back up your servers in Azure but not turn them on. This way, you’re not paying to keep those VMs running constantly. You’ll only pay for them in that worst-case scenario when you’ll rely on them to keep operating.
But there’s another caveat here. Many companies don’t really ask themselves what systems are essential in a disaster recovery scenario. This leads to making everything available in the cloud in a disaster scenario, which gets very expensive. Without strategically choosing which systems to make available, then modeling the costs of those systems, you can find a surprising bill for cloud systems in a disaster recovery scenario.
3. Consider repatriating some systems to hardware (i.e. adopting a hybrid cloud architecture)
In some cases, organizations may migrate all systems to the cloud without fully understanding the cost. As time goes on, cloud delivery becomes unsustainable for some systems. If a problematic system can’t move from a server model to a services model, or if services don’t work either, then it may make sense to “repatriate” the system to hardware—i.e. undergo a reverse cloud migration.
Typically, that means moving from a 100% cloud architecture to a hybrid cloud model—one in which you host some systems on premises (or in the private cloud) and some systems in the public cloud. While hybrid cloud management comes with its own challenges, it offers certain benefits when applied strategically.
Repatriation isn’t right in every scenario, but the key is to know your needs and capabilities inside and out. Read more here: Cloud Repatriation 101: What’s Right For You? Ultimately, if you’re not sure what’s best for your organization, an expert cloud consultant can help.
4. Review cloud costs on a regular basis
Cloud cost optimization isn’t a “set it and forget it” initiative. Vendors like Amazon, Microsoft, and Google will adjust their pricing periodically. Your needs may fluctuate too, which changes the calculation of how best to fulfill those needs.
Cost optimization doesn’t require continuous attention, but it does require periodic attention. The best plan is to do it monthly or quarterly, depending on how frequently your needs change—and how tight your margins are.
Of course, many midmarket companies already struggle to fulfill their IT obligations. Sometimes cloud cost optimization is simply too much net-new responsibility for internal staff. In these cases, a managed service provider with cloud expertise can help.
Cost optimization in a multi-cloud scenario
If you’re working with multiple public cloud vendors, cost management gets far more complex—so much so that a multi-cloud approach rarely makes sense for midmarket companies. While solutions exist to pull all your cloud management information into a single pane of glass, it’s just too much heavy lifting for organizations with limited IT resources.
Not only is multi-cloud challenging, it’s rarely worth it for midmarket companies. Even if one vendor offers you cheaper email hosting than your primary cloud vendor, the increased complexity of management introduces hidden costs. It’s almost impossible to calculate whether that “cheaper” email hosting will actually cost less in real life. Rather than trying to optimize in this scenario, it’s best for midmarket companies to plan strategically so they avoid this scenario.
Cost optimization in a hybrid cloud scenario
If it’s done right, hybrid cloud comes with cost optimization built in.
You bought hardware strategically to take advantage of the capex (capital expense) cost model as appropriate. The longer you run that hardware, the cheaper your unit cost of processing data becomes.
Of course, bare hardware without any ongoing licensing would be the cheapest—but this isn’t realistic. You’ll have some form of opex (operating expenses) alongside the capex component. These ongoing costs take the form of OEM vendor support, warranty contracts, and so on.
Of course, owning your own hardware means you’re responsible for licensing your operating system. You also have to license your hypervisor (the software that allows you to run multiple virtual machines on one physical machine), operating systems, and so on.
Outside of those things, you really have to get strategic with licensing. In fact, deciding where to license an application (on premises or in the cloud) is the key to cost optimization in a hybrid scenario. Maybe you have Microsoft 365 licenses coupled with VMware, or maybe you have Windows Server licensing too. Whatever the mix, you’re going to see some opex in addition to the capex. Hybrid cost optimization is all about hitting the ideal balance.
The takeaway: Get your strategy right, and cost optimization will follow
While you can always address cloud cost issues after an implementation, it’s less painful to start with a solid strategy and purchase the cloud services that support that strategy. This helps you avoid further migrations (or reverse migrations).
Some organizations have the expertise on staff to do this well. Others have their hands full with day-to-day IT operations. In these cases, a strategic cloud services provider can help. Here at Corsica Technologies, we know cloud cost optimization inside and out. Our experts can help you audit your functional needs, align with the best offerings of the market, and make sure you’re getting the most out of the cloud.
Want to learn more about cloud cost optimization?
Reach out to schedule a consultation with our cloud specialists.
