For years, cloud egress fees – the costs incurred when data moves out of a cloud environment – were like death and taxes. They were unavoidable for any organization that used the cloud. But that has changed over the past few months as major cloud providers Google, Amazon and Microsoft have announced the discontinuation of egress fees, at least under certain circumstances.
This sounds perfect, right? You can now use the cloud when it is appropriate, and when it is not, you can move your data back without a penalty. However, most organizations will find that the removal of egress fees does not translate to storage cost savings.
Why? There are a few reasons, but perhaps the main one involves the nature of data itself. Fundamentally, the use and value of data shifts throughout its lifecycle; optimizing its costs requires dynamic data placement and dynamic right-sizing of data storage and protection to optimize costs. Without this ongoing data management, any moves to the cloud or back from the cloud may temporarily alleviate some costs but will never provide sustainable savings. In fact, the cloud offers a wide variety of storage tiers and storage classes that, when leveraged correctly, should offer significant savings.
Changes to Cloud Egress Fees Explained
Since the dawn of cloud computing, most cloud providers charged egress fees whenever customers moved data into a different cloud or to an on-premises environment. There were exceptions for certain types of data, and some cloud providers allowed a limited amount of data egress each month at no cost. But in general, whenever you wanted to move data out of a cloud environment, you paid a fee for every gigabyte of information you transferred.
This changed in January 2024, when Google Cloud announced that it would no longer charge for data egress when customers migrated out of its cloud platform. Amazon Web Services (AWS) and Microsoft Azure followed suit in March.
The change in policy – which arguably has at least as much to do with complying with regulations such as the European Data Act as it does with benefiting customers – doesn’t eliminate cloud egress fees entirely. I’ll explain this in more detail below.
Still, this is, in some respects, a momentous update because it is now considerably cheaper for organizations to move workloads out of the cloud and to new environments. Gone are the days when the cost of data egress was a roadblock that prevented migrations, even in situations where there was a clear rationale for switching to a different cloud or repatriating workloads back on-prem.
The Caveats Surrounding Egress Fee Updates
In other ways, the elimination of egress fees feels less earth-shattering. The biggest limitation is that Amazon, Microsoft, and Google offer free egress only if you migrate off their platforms entirely. They’ll still bill you if you move data out of their platform, and then move it back later.
Another caveat is that in some cases, the elimination of egress fees applies only to data transferred out of standard cloud storage services. For example, Microsoft states that data transfer charges for “specialized services including Express Route, Express Route Direct, VPN, Azure Front Door, and Azure Content Delivery Network (CDN)” still apply.
What’s more, egress fees typically account for a relatively small portion of overall spending related to cloud data. The complex pricing schedules for cloud services mean that your total bill will also reflect data storage charges, charges for accessing or changing data, charges for logging and monitoring, and on and on.
In this sense, no longer having to pay egress fees is like getting drip coffee for free but still being charged for the cream, the sugar, and the cup that holds it all. It may save you a bit of money, but the savings are limited, especially if you don’t carefully manage other costs.
What It Really Takes to Optimize Cloud Data Costs
Organizations that really want to save money in the cloud should take an analytics-based approach to understanding their data assets. This allows IT to make nuanced decisions on data storage and data management rather than jumping on the latest cloud pricing changes to move into a different platform or back in-house. Moving data every year or two from one platform or cloud service to another in a quest for the best deal is a wild goose chase that never delivers the desired results and consumes massive resources and time from overstretched IT teams.
Consider that you could achieve a better financial outcome by keeping your data in the cloud where it already lives, while making changes to your storage configuration within that cloud. For instance, you could use data management to dynamically tier cold data to lower-cost storage tiers, such as from higher-cost AWS FSx to lower-cost AWS S3 Glacier.
Or perhaps you could assess your data assets and policies and delete cold or old data that you no longer need to reduce your overall data volume and, by extension, your storage costs. Before you move any data from on-premises to the cloud, be sure to understand what data is active or hot, what is warm, what is inactive or cold, and what requires extra security so that you are right-placing it in the appropriate cloud storage tier.
Consider, too, your longer-term plans when developing a cloud strategy: whether they include AI investments, accelerated data growth predictions or potentially acquiring a company that has data and storage to integrate and consolidate, for instance. Make sure you choose a data management and data storage platform that allows for flexibility. For example, unstructured data management solutions that support cloud-native access to data ensure that you will not be locked into the original storage vendor’s platform and you can access and move your data without licensing penalties.
Conclusion: A Balanced Approach to Cloud Migration
The partial elimination of cloud egress fees makes it easier than ever to migrate workloads and data to different platforms. But that doesn’t necessarily mean that migrating data to another cloud or back on-premises will be cheaper in the long run.
The best way to optimize your cloud strategy is to shift your perspective from managing storage to managing data. Knowing which data you own, understanding data growth patterns, what data is hot versus warm versus cold, and factoring in longer-term plans for AI and data reuse are necessary to create the right plan. Use analytics to determine which cloud storage solutions best meet the needs of your various data assets today and down the road. Ideally, you can use analytics and policy-driven automation to move data at the right time to the best storage tier –without excessive storage licensing, egress, and other penalties. This approach will save you the most in the long term while also satisfying the diverse needs of users and meeting corporate requirements.