Google Cloud just eliminated data transfer fees. Here’s why.

Earlier this month, Google announced that it would remove data transfer fees for those looking to move their data off of Google Cloud.

It's an interesting move that no major cloud provider has made before, and I'd like to talk about why it happened now, and what it means for the future of cloud computing.

Let's dive in. 

Unpacking Google's announcement#

On January 11, Google Cloud announced, effectively immediately, that they would be removing data transfer fees for customers looking to migrate their data off of Google Cloud. 

In removing data transfer fees, Google is empowering customers who wish to migrate their data to another cloud provider and/or on-premise solution. By allowing customers to scale in a way that best suits their situation, Google emphasizes they are upholding the "promise of the cloud," with their customers' best interests in mind.

Interestingly, their announcement on the Google Cloud Blog also makes a concerted point to condemn the actions of other major cloud providers.

"Eliminating data transfer fees for switching cloud providers will make it easier for customers to change their cloud provider; however, it does not solve the fundamental issue that prevents many customers from working with their preferred cloud provider in the first place: restrictive and unfair licensing practices."

Amit Zavery, GM/VP, Head of Platform, Google Cloud

Google is attempting to distinguish themselves from the "unfair" practices of what they call “certain legacy providers." Google certainly has their competitors Amazon and Microsoft in mind with these statements, and links directly to articles about their practices. (It's worth noting here that Google Cloud ranks third in market share behind AWS and Azure — we’ll discuss more about Google’s likely business strategy in a moment).

The significance of free data transfers#

Most cloud platforms will charge a lot of money to move data to another platform, so Google is taking substantial initiative to make this free for customers.

A lot of providers, like AWS and Azure, offer free ingestion of data. However, transferring or removing data outside of their network can add up to be costly.

For example, let's consider Amazon S3. Each month, they allow 100GB of data transfer for free. But above that, you'll have to start paying. One estimate shows that the cost for transferring 50TB of data between AWS via the public internet would be around $4,491 per month. 

Data transfer costs will add up depending on the amount of data you're trying to move. With terabytes of data, it ends up being cost-prohibitive to move to another platform. 

But of course, Google is not saying that all data transfers will be free. Applications that remain deployed on GCP will not be getting free or cheaper bandwidth. 

Generally, data transfer costs are higher when they occur between different cloud providers, and lower when they're within the same provider's network. This status quo is still going to be the case for those remaining on Google Cloud. What's changing is that Google is no longer going to make it cost-prohibitive for users to migrate off their cloud. 

Vendor lock-in vs Interoperability#

In their announcement, Google seized the opportunity to make a dig against their competitors for engaging in the age-old battle of vendor lock-in instead of having more interoperable platforms.

Interoperability is the ability of different cloud services and platforms to seamlessly work together. Achieving interoperability requires various practices, including having standardized protocols and compatible interfaces, and ensuring efficient data exchange.

Interoperability is also essential for avoiding vendor lock-in, which is when customers are disincentivized from working with multiple vendors. Unfortunately, current pricing systems do encourage vendor lock-in.

Google's blog post notes one of Microsoft's licensing practices, which involves forcing Amazon users to pay more for the same services. 

Google specifically calls out the fact that Microsoft charges users 5x more to run Windows Server and SQL Server on AWS than on Azure.

Similarly, if a customer wants to run a certain Windows desktop app on AWS, they have to pay more than what they would if they run it on Azure. If they ran it on Azure, they would get a subsidized license, but on AWS they'd have to pay more (i.e., they'd have to pay the OS and product licensing fees, in addition to infrastructure pay-as-you-go costs).

The complexity of interoperability worsens with the fact that Microsoft prohibits Amazon from buying their licenses at a discount. Rather, every customer must bring their own license (this is called BYOL). If a customer wants to run Microsoft software on AWS, Microsoft allows it, but will only deal with customers individually on a BYOL basis. 

Google is taking a step in the opposite direction here. But they're also in third place in the cloud market, so this move won't hurt them that much. If anything, they may be hoping this move will push other vendors to do the same. After all, if AWS and Azure allowed cheaper data transfer, then Google Cloud could gain new customers themselves.

Interoperability in multi-cloud strategies#

Interoperability is essential for a more open and flexible multi-cloud strategy. A multi-cloud strategy involves using services from multiple cloud providers. This can be beneficial for various reasons, giving organizations the flexibility to combine specific services that best suit their needs, prepare for disaster recovery, and save money by choosing services from vendors that offered them a better deal.

In order to allow interoperability, cloud providers must make it easy to transfer workloads and data across different clouds.

Technically, all cloud platforms claim to be interoperable, and say that you can easily transfer your data.

In reality, interoperability is the biggest hurdle in a company's multi-cloud strategy. However, there are some cases where interoperability doesn't pose a huge challenge. 

Let's consider Apple, for example. Let's say Apple's customers have iCloud backups spread across AWS and Azure. In this setup, iCloud backups are handled independently by each respective cloud provider. There's no need for complex interactions between AWS and Azure here, so interoperability doesn't really pose a problem here.

Yet in other cases, achieving interoperability is more complex. 

The challenge intensifies, for instance, when an app runs on one cloud provider, but the data it requires resides elsewhere. Running on the same vendor's cloud incurs negligible network costs due to favorable pricing. However, when an app like Notion runs on AWS, and user data resides on Azure, costs escalate. AWS charges more for outbound network calls, and Azure charges more for data retrieval from an external network. 

So, while multi-cloud interoperability exists in theory, implementing it is far from straightforward. It can also be cost-prohibitive.

As a result, many large companies end up running their complete services with one vendor. The assumption is that these isolated services do not have to communicate frequently. For instance, an organization might have a multi-cloud deployment where they use different public clouds for different environments (e.g., staging on Azure, development on AWS, and deployment on Google Cloud). When these services need to communicate, the transfers would be so small and infrequent that the cost is not that high and is therefore acceptable. 

"At Google Cloud, we work to support a thriving cloud ecosystem that is open, secure, and interoperable." — Amit Zavery, GM/VP, Head of Platform, Google Cloud

Google positions themselves as advocates of interoperability. At the end of the day, all cloud platforms charge more to transfer data externally — Google Cloud included. So, although Google wants to distinguish themselves to be enablers of a flexible cloud ecosystem, they too participate in practices that can be cost-prohibitive.

What does Google have to gain?#

At first blush, it might seem to be a bit of a head-scratching move that Google is making it easier for customers to leave their platform. Google certainly is taking a step in the right direction (and away from vendor lock-in), but companies often do this because it's self-serving.

It seems likely that they are taking this step now in order to make some noise and force movement in what has become a rigid market. After all, let's remember that Google is in third place in terms of market share and popularity in the cloud market right now. (Just take a look at Stack Overflow's breakdown of popular cloud providers in their most recent Developer Survey.)

While Google Cloud doesn't have a lot to lose with this change given their current standing, it does come off as a gesture of goodwill that could lead to good press (and the tone of their blog announcement certainly reflects such intentions — at least in how they are positioning against the competition). 

If Google Cloud isn't the optimal choice for your needs, Google is choosing not to create additional barriers to prevent you from moving on. For instance, let's say that Google Cloud can't meet your regulator's requirements for data placement. If Google doesn't have a data center in the required geographic boundaries, then Google Cloud may not be able to suit your needs. 

There may be other use cases for which Google Cloud isn't optimally designed, and allowing migration off their platform can also be a good way for Google to part ways with holdouts that require heavy investment to maintain compatibility. 

What to expect for the future of cloud#

So, how will this all shake out? Will AWS or Azure follow suit? I would be shocked, if so.

If either Azure or AWS made a similar move, then the other counterpart would be forced to do the same. This move would certainly be unfavorable for AWS, who is still dominating cloud market share. Since Microsoft is making significant progress on AI, there is actually an incentive to move to Azure from AWS, and we would likely see a large migration to Azure if AWS made it easier for users to leave their platform.

As long as the market sees healthy competition, we will continue seeing more changes among cloud providers for years to come. While I've said that Google Cloud is a distant third in the top cloud platform providers, they'll be around to stay. They're making significant investments to improve infrastructure, and have quite a large amount of satisfied customers. 

If you're a developer who hasn't yet been working with cloud services, I'd recommend that you do so this year — especially because we're now in a time of major flux stemming from the Generative AI paradigm shift. The spread of AI is expanding our cloud services and challenging our infrastructure on an unprecedented scale. Looking ahead, we'll need developers with cloud skills as we continue to reimagine and rebuild our applications and cloud infrastructure in alignment with AI trends. 

If you're not sure which platform to start with, AWS is a good place to start. As the dominant cloud provider, you're very likely to run into AWS at work. You can get hands-on with and work directly with AWS services without any setup or cleanup with Cloud Labs at Educative.

Happy learning!