There’s no single formula for calculating your hybrid cloud TCO, but here are all the basic components that should go into that calculation.
Managing cloud costs is hard enough when you use a straightforward public cloud architecture. It gets even harder when you move to a hybrid environment that integrates public cloud services with private infrastructure. And don’t forget to add data center colocation to the mix.
With that reality in mind, here’s a guide to calculating total cost of ownership (TCO) for hybrid cloud architectures that run inside a colocation data center.
Hybrid Cloud Software Pricing
The most obvious expense associated with a colocated hybrid cloud is the cost of the software that you use to build and manage your cloud.
For many organizations today, that software will come in the form of a platform like AWS Outposts, Azure Stack, or Google Anthos, which make it possible to run public cloud services and management tools on private infrastructure.
At a basic level, each of these services uses the same pricing structure. The vendors charge based on the number of virtual CPUs (vCPUs) that customers run within their hybrid environment. Outposts is somewhat different in that it is priced based on compute instance types, but this is more or less a proxy for vCPUs.
If you use Anthos and Outposts, you can save some money by paying upfront or committing to a monthly subscription. Azure Stack pricing includes only a pay-as-you-go option.
Public Cloud Egress, API, and Storage Fees
A less obvious cost associated with hybrid cloud platforms like those described above are the extra fees you’ll pay for interactions between your private infrastructure and the public cloud.
Those fees include things like data egress charges that public cloud vendors assess when you move data from their clouds into your own storage media. They generally apply even if you are using an environment based on a platform like Azure Stack or Outposts. API calls to public cloud storage services usually incur a fee, too.
On top of this, some hybrid cloud services charge separate fees even for storage that you host yourself. Azure charges fees for managing your disks within a hybrid cloud, for example.
It’s easy to overlook the costs of things like data egress and API fees in the public cloud. It may be even easier in hybrid environments, where you might assume that these fees are built into the basic cost of the hybrid cloud software that you use. Generally, they’re not.
The cost of the servers that you use to host your hybrid cloud is another significant factor in your hybrid cloud TCO.
If you use Azure Stack, Anthos, or most other hybrid cloud platforms (like Eucalyptus), you’ll need to supply your own servers. The cost of doing so with Azure Stack is likely to be higher than with other platforms, because Azure Stack works only with certified hardware. That means users may not be able to use servers they already own to build a hybrid cloud based on Azure Stack. It also means they will have fewer purchasing options. Anthos and Eucalyptus aren’t subject to these restrictions; they work with any type of modern server.
Hardware costs for AWS Outposts are bundled into the cost of the Outposts platform, because AWS supplies the servers (which is why Outposts costs thousands of dollars per month for each server, whereas the other hybrid cloud platforms charge only dollars per month per vCPU). This makes hardware costs for AWS more straightforward and less variable. They may be high, but at least you know exactly what you’re going to pay for hardware before you commit to Outposts.
Colocation and Interconnection Costs
When you run a hybrid cloud inside a colocation data center, colocation costs are another key expense that contributes to your TCO.
Calculating these costs can be difficult, because the specifics of colocation pricing vary from provider to provider. You may need to pay for resources like electricity and network service as separate costs, or they may be built into your colocation bundle.
You will also pay your colocation provider for the network links between your and your cloud provider’s infrastructure. Those costs alone are almost never a straight-forward calculation.
Thus, there’s no simple way to determine how much colocation will add to your hybrid cloud TCO. But whatever the details of your colocation plan, the costs are likely to be significant, so you’ll want to assess them carefully before committing to a colocated hybrid cloud.
Management and Support Costs
The final factor to consider in hybrid cloud TCO is the cost of deploying, managing, and supporting your hybrid cloud environment.
These expenses will vary depending on which platform you use. They’re likely to be lowest in the case of AWS Outposts, which is a fully managed service, with minimal deployment or upkeep effort required on the part of customers.
Azure Stack and Anthos leave more up to the user. That said, the fact that these platforms for the most part use the same management tooling as the public clouds with which they are associated means that, if you already know how to use those tools, you won’t face a steep learning curve when adjusting to hybrid cloud management.
One advantage of using colocation data centers to host your hybrid cloud is that you may also be able to obtain management and support services for the hybrid environment from the colocation provider. AppScale, which sells a hybrid cloud framework based on Eucalyptus, is partnering with some colo providers around support services, for example. But for now, colo packages that bundle hybrid cloud management with colo space and services are the exception.
When it comes time to determine how much a hybrid cloud will cost you, there are a variety of factors to consider. Hybrid cloud software and infrastructure are the most obvious, but it’s critical to include several other types of expenses as well when calculating hybrid cloud TCO.