For more expert advice on optimizing cloud computing costs, read our Cloud Cost Optimization: Best Practices for IT Pros report.
Moving data to the cloud can be a great way to save money. With costs that are measured in mere pennies per gigabyte, cloud storage may be much less expensive--not to mention more reliable and easier to manage--than on-premise storage infrastructure. Yet those pennies-per-gigabyte add up over time. If you’re not careful to control your cloud storage costs, you may end with a total storage bill that far exceeds what you were hoping for.
With that challenge in mind, here’s a rundown of best practices for minimizing cloud storage costs without sacrificing storage reliability and performance.
1. Choose a low-cost cloud storage service.
Probably the best way to lower your total cost of cloud storage is to make sure you choose the lowest-cost storage platform from the get-go.
In many cases, this will mean choosing an “alternative” cloud provider, rather than one of the “Big Three” public clouds. Amazon Web Services (AWS), Microsoft Azure and Google Cloud Platform (GCP) all offer affordable storage services, but you are likely to find better pricing from cloud vendors that specialize in storage. Wasabi and Backblaze, for example, boast storage prices that are a fraction of those on the larger clouds.
The caveat here is that alternative cloud storage providers may not offer all of the features you need. They typically provide only object storage that is equivalent to services like AWS S3 and Azure Blob Storage. They don’t usually offer block storage or the wide selection of databases that you can find on a Big Three cloud platform, nor do they offer other types of cloud services, like virtual machines, that you may want to run in the same cloud where you store your data.
The bottom line: If you just need object storage, be sure to assess the services and pricing of smaller-name cloud providers. Depending on your needs and use cases, these vendors may or may not be able to deliver the storage you need at a significantly lower cost than AWS, Azure or GCP.
2. Choose the right storage service.
If you do opt to store data in a cloud that offers multiple types of storage services, be sure you choose the most cost-effective storage type for your needs.
Typically, object storage is the simplest and most economical way to store data in the cloud. But, depending on what you’re doing with the data, you may achieve a lower total cost by using a different storage option. For example, data that needs to be processed in a certain way may be better stored in a database, where it can be structured such that the time (and, by extension, money) your applications spend ingesting and analyzing it is reduced. With object storage, you have little ability to structure data, which can make analysis more expensive.
3. Take advantage of storage classes or tiers.
On many public cloud platforms, object storage services are available in different “classes” (the term AWS and GCP use) or “tiers” (Azure’s terminology). The standard storage class or tier is the most expensive. You can save money by choosing to store data in a lower-cost tier. The concept here is quite similar to on-premises storage tiering strategies.
The tradeoff is usually that the lower-cost storage options offer lower levels of performance. Data stored on these tiers is not accessible instantaneously; it can take anywhere from several minutes to several hours to read it. You may also have to pay extra fees if you access the data on a frequent basis.
Depending on what you’re doing with your data, then, lower-cost storage tiers or classes may or may not be a good way to lower your cloud storage spend. Archival data that you don’t expect to access regularly is a great candidate for low-cost tiers. Data that your applications must read or write on a frequent basis is not.
4. Use cloud data lifecycle policies.
What if you need your data to be in a standard storage class sometimes, but not other times? That’s where data lifecycle policies come in. On most major public clouds, you can write policies that automatically migrate data from one storage class or tier to another. These policies save you money by ensuring that you don’t keep data in a higher-cost class longer than you need to.
Lifecycle policies are a great way to lower your cloud storage costs if, for example, you have log data that you initially need to keep accessible in case you have to refer to it for troubleshooting purposes. Once the logs get older and are less likely to be needed, you can move them to a lower-cost tier automatically using a lifecycle policy.
5. Be smart about data reliability.
Protecting data against temporary unavailability or permanent loss is typically a priority for any enterprise. It may even be one of the reasons why you move data to the cloud in the first place: You believe the cloud will prove more reliable than your on-premises infrastructure.
That said, it’s important to balance data reliability with cost. The default levels of availability that most cloud providers are pretty good; Amazon, for example, famously promises “11 9s” of durability for data stored in its S3 service. If you’re comfortable with a 0.000000001 percent chance of data loss, and you trust Amazon to live up to its promises, the default level of durability in S3 should be good enough for you.
On the other hand, you may need even more reliability. In that case, you could choose to mirror your data across multiple cloud availability zones or regions, or even store redundant copies in different clouds. Doing so will increase the chances that your data remains safe and available in the event of a cloud infrastructure failure. But it will also substantially add to your storage costs.
Whichever approach you take, be sure to do a cost-benefit analysis to make sure the expense is worth it. No matter how much you spend on cloud storage, you can never completely, absolutely guarantee 100% availability.
6. Don’t store more data than you need.
Lastly, you can minimize cloud storage costs by making sure you only store data in the cloud if you actually need to.
This point should be obvious, but given just how easy it is to upload data to the cloud--and how hard it is to keep track of it once it’s there--you may inadvertently end up with bloated cloud environments filled with data that you no longer need to retain or that should be moved back on-premises.
To protect against risks like these, make sure you understand the architecture of your cloud storage strategy, and that you know which data lives where. You can also use the tagging features that are built into most cloud storage services to label data and make it easier to find. For instance, if you have to keep some data around for a certain length of time for compliance reasons, you can use tags to label that data and then delete it after the mandatory retention period has expired.
Cloud storage can be dirt-cheap, but only if you approach it intelligently. Be sure to choose the right cloud platform, the right storage services and the right data management strategies to keep your cloud storage costs as low as possible.