Cloud computing seems like an offer you can’t refuse. Instead of buying your own equipment and taking the time to set it all up, you can rent it from a cloud service provider and pay as you go. Even better, you can use only as much as you like, scaling up computing and storage resources when you need them.
Like many offers you can’t refuse, though, there is a hidden catch. Private cloud computing costs can mount up over time, leaving you with a hefty bill at the end of each month.
Cloud management software company SoftwareONE surveyed 300 IT decision-makers last year to gauge their experiences with the cloud. The biggest pain point, felt by 37%, was unpredictable budgeting. 26% felt that cloud pricing models were more complex than on-premises ones.
Here are some tips you can use to ensure that your adventures in the cloud don’t leave you flat broke:
Assess and price all services
Be sure you understand all the services you’ll be using when moving your services to the cloud. Platform as a service (PaaS) implementations offer feature additional functions like identity and access management or database services separately, with individual pricing.
Optimize your applications and infrastructure for cloud environments
Design your cloud-based infrastructure with an understanding of data transfer fees, as these can quickly add up. For example, it’s free to get external data into Amazon Web Services, but it costs to get it out to the Internet. This will affect your hybrid or multi-cloud financial models.
Similarly, Amazon charges to send traffic from one AWS region to another, and it charges to send and receive information between availability zones in the same region. Factor these costs into your cloud-based infrastructure design.
Use your management console
What’s worse than a costly virtual machine? A costly virtual machine that no one is using. Yet companies regularly leave cloud resources running, racking up unnecessary costs. Respondents to Flexera’s 2019 RightScale State of the Cloud report estimated 27% waste, while Flexera measured actual waste at 35%.
The answer is to use automated policies that shut down unused machines and rightsize virtual instances. Few of the survey respondents did this, and it’s an area where you can excel.
Classify your data
Data storage is another area where costs can quickly mount. Don‘t leave data in the cloud that doesn’t belong there, or to store data on an inappropriate class of cloud storage. This can quickly drain your budget.
Classifying your data will help you store it in the right place, such as a single-zone storage area rather than a multi-zone redundant storage configuration, or in an archive rather than in primary storage. You may even find that some data needn’t be in the cloud at all.
Use reserved instances where possible
Sure, the marketing rhetoric says that the cloud gives you unlimited, on-demand computing power, where and when you need it. That’s true, but it comes at a price. If you can plan your computing use, then you can score big discounts by using reserved instances.
Reserved instances require you to pay up front for the cloud computing resources you think you’ll need, rather than buying them on demand and paying as you go. They come with substantial potential savings of up to 65% by some estimates, depending on the instance you’re using and the length of the contract.
The catch here is that you have to plan well. Get it right, and you could save big. Get it wrong, and a poorly planned contract could drain your budget.
These pointers will help you tame your cloud budget before it gets out of control.
Considering them early will help you build those savings into your long-term cloud operation.