NetApp AltaVault is a cloud-integrated storage solution that can automatically back up and archive your on-premises data to dozens of leading cloud services, including AWS, Azure, and Google Cloud Platform. We’ve seen tremendous growth in the adoption of cloud-based backup as our customers integrate cloud services into their IT operations—a 175% year-over-year increase in cloud storage capacity managed by AltaVault appliances as of the last count.
AltaVault has been so successful because it helps our customers respond to the rapidly evolving market for cloud services so they can reap the benefits of cloud elasticity, price, and availability. For example, the AWS cloud service offers multiple storage classes, each with different price and performance characteristics. When backing up to AWS, AltaVault users can select among three storage classes:
- Amazon Simple Storage Service (S3). The original AWS massively scalable object store
- S3 Standard – Infrequent Access (S3 IA). Provides many of the features of S3, but at a lower price point
- Amazon Glacier. Extremely low-cost cloud storage service for data archiving and long-term backup, with access times to data that range from minutes to hours
With three options to choose from, I wanted to find out which AWS storage class AltaVault customers prefer for their backup data, and why. I realized that I could get answers to these questions by looking at NetApp AutoSupport data that aggregates the cloud service selections made by a large population of AltaVault users over time.
Cost Is King When Backing Up to the Cloud
It turns out that more than 70% of AltaVault customers who enable our remote AutoSupport service are using AWS as their cloud backup target. This shouldn’t be surprising. AWS is the industry’s leading provider of cloud services, and our customers overwhelmingly choose it because it offers a range of options, as well as availability and cost effectiveness.
Looking across the three AWS storage classes available to AltaVault users, I discovered that Amazon Glacier is by far the most selected storage type, with more than twice as much data stored in Glacier as in S3 and S3 IA combined. This data point is intriguing. Given that the majority of AltaVault data storage represents backup use cases rather than archival use cases, why is Glacier the most popular cloud target? Shouldn’t more users be selecting S3 or S3 IA, given the faster access they provide when customers need to recover data that is critical to meet business needs?
There are probably many reasons behind the selection of Glacier, but the top three that come to my mind are:
- Acceptance of cloud solutions is often determined by cloud price sensitivity.
- Glacier offers the lowest costs among the three AWS storage classes—about 5x lower than S3 and about 3x lower than S3 IA.
- Glacier was introduced in 2012, just as AltaVault adoption was accelerating.
By looking closely at the costs for each storage class over time, we can get a better sense of the importance of storage costs when customers select their cloud backup targets. The following table shows the estimated average monthly costs of sending and storing 100TB of weekly backup data to S3, S3 IA, and Glacier, calculated over a 3-year period. This example tells me that when it comes time for AltaVault users to select a cloud backup target, cost is king.
Potential Trade-Offs When Using Glacier
Although Glacier is a perfectly acceptable storage class for a backup use case with AltaVault, I would be doing a disservice if I didn’t point out some of the trade-offs you might encounter when using it as a backup target. There are three trade-offs to be aware of:
- Unanticipated costs. When your backup data needs to be restored on a frequent basis, using Glacier could increase your recovery costs. If most of your data can be recovered from the local AltaVault cache, the good news is that your costs for recalling data from Glacier will be minimal. However, if you’re regularly recovering data from Glacier because it has aged out of your local cache, this could quickly eat up your cost savings from using Glacier, and then some.
- Performance. Glacier is Amazon’s slowest response storage tier, which increases recovery times—it typically takes several hours before Glacier can begin serving data for a restore request. If SLAs and RTO are important, you should consider using a higher-performing AWS storage class.
- Ability to automate. Using Glacier as an object storage target introduces manual operations for AltaVault, such as the need to prepopulate data to the AltaVault cache before it can be restored. This reduces the automation benefits that AltaVault can provide as a replacement for traditional tape backup and archives. If the administration of your cloud backup solution must be automated, Glacier won’t work for you.
Fortunately, the AWS S3 IA option can address all these issues. The capabilities of S3 IA are almost identical to those of S3, and it has a more attractive price point, so I expect AltaVault users to increase their adoption of S3 IA over time. However, as the lowest-cost AWS storage class, Glacier will probably continue to be a popular choice—assuming that users choose to accept the trade-offs that go along with the lower costs.
In summary, AltaVault offers customers long-term, cost-effective storage for protecting backup and archive workloads. As companies gain more experience with their cloud options, reducing the cost of storage is likely to become just one of many considerations that are evaluated when selecting a backup target. Requirements for retention, accessibility, and longevity should also be part of any cloud data storage decision.
Try AltaVault at No Charge
To learn first-hand the advantages of efficient backup and recovery with NetApp AltaVault cloud-integrated storage, download the software-only version and start your free 90-day trial today.