Trends pile on top of trends. As the world and the people in it move faster with an “instant gratification” mindset, there’s an even bigger impact on how engineers evaluate products, tools, and implementation techniques.
Cloud, on-prem, and “legacy” systems continue to thrive and have a critical impact. The question now becomes “How can we make them easier to manage?”.
In this blog post, you’ll learn about why hybrid cloud is important and a few solutions to help you on the journey.
Not Everyone Is Moving To The Cloud
Currently, there are 5,390 data centers across America.
10 of them are Microsoft Azure.
The exact number for AWS is disclosed (but they have 100 across the globe, not just in the US).
GCP has 40 regions globally.
Doing the math there, it’s relatively safe to say from an assumption perspective that there are 5,000 data centers that have nothing to do with the major cloud providers. These are standalone data centers where organizations can buy/host racks, smaller public clouds, and individual enterprise data centers.
The point is that’s a lot of workloads on-prem.
Despite what everyone thought the cloud would become (there was a lot of talk about nothing being on-prem anymore), some workloads still run on-prem for a variety of reasons.
With that being said, it’s safe to say that for the foreseeable future, infrastructure on-prem that’s managed outside of cloud providers still matters.
Workloads Are Still On-Prem
Now that we have a fair guestimate about how many data centers still exist for workloads that are outside of the major cloud providers, let’s talk about a few use cases.
Compliance
There are a lot of compliance needs from healthcare to finance to government sectors and everything in between. In short, it typically comes down to where the data can legally or ethically exist. For example, there are certain HIPPA and GDPR regulations that state certain types of data must be able to be managed 100% without any abstractions that the cloud gives you.
Edge
I have two colleagues that I’ve spoken to with some pretty interesting edge device stories.
The first is edge devices that have to quickly connect to fighter jets for updates (for obvious reasons, they can’t be public facing) and those devices need to be set in certain geographical areas.
The second is for mines (as in, mining work). Devices simply cannot connect all the way down in a mine, so they need edge devices to perform various activities for them in terms of necessary connections while in the mines.
A third example is Chick-fil-A, a popular (and quite tasty) fast food restaurant that’s known for chicken sandwiches (and tasty fries… and tasty shakes… and tasty everything). Each Chick-fil-A restaurant has an Intel NUC (or a machine like a NUC) running the cash register software. They’re all managed by the Information Technology (IT) or remotely.
As you can see, Edge devices could be needed for anything from hard-to-reach areas to sensitive practices/implementations to tasty treats.
Performance
When it comes to the need for speed and overall software endurance, performance could still potentially be an issue in the cloud depending on regions and bandwidth pipe capabilities. It could be anything from latency issues to having dedicated resources to overall predictability. If you’re managing the infrastructure, it’s far easier to predict when resources will be available and how much bandwidth you have left in comparison to renting it from cloud providers.
The Ability To Work Wherever On Whatever
On-prem is still needed. Bare metal is still needed. However, a lot of those organizations using on-prem implementations also use the public cloud in some capacity.
For example:
- AKS on-prem
- AKS in the cloud,
- container services
There also needs to be a way to manage VMs in the same way. Having workloads that run on-prem and for scalability as an “emergency” of sorts that can “burst” into the cloud makes a lot of sense. If there’s an issue in a data center or an environmental catastrophe or even something like “the servers are delayed and won’t be here for a week”, organizations can use bursting capabilities to send workloads to the cloud.
Engineers need the ability to work on the environments they’re implementing the same way regardless of where the infrastructure exists.
Enough time has gone on for cloud providers and enterprises/organizations to realize that not every single workload will be in the cloud. There needs to be a way to manage all workloads, wherever they exist, in one location.
💡
A hybrid solution is great for a lot of organizations, but it’s not necessary. As a good example here, if you’re a startup building an enterprise application, chances are utilizing the public cloud is going to be much more efficient than hosting in a data center until it’s needed. Just keep in mind that, especially if you’re a consultant, not every organization is going to need this solution.
Hybrid Solutions
To manage workloads not only in one location but in the same way, hybrid solutions can be implemented. Luckily, the majority of hybrid solutions very much have the capabilities to make it look like all of the workloads are running in the same place. For example, you could be managing Kubernetes clusters via one API and one UI and they’re in different locations like on-prem and in the cloud at the same time.
Let’s discuss a few tools that do this.
Azure Local
Formerly called Azure Stack HCI, Azure Local gives you the ability to manage VMs, Azure Kubernetes Service (AKS) clusters, and even various services like Azure Functions in your own data center while managing it all within the Azure portal. It’s an Operating System that runs on your bare-metal infrastructure which has the capability to connect to the cloud via encryption methodologies for authentication/authorization.
You could even test this out by running Azure Local on an Intel NUC and managing/deploying workloads to the NUC sitting on your desk from the Azure portal.
EKS Hybrid
Hybrid within AWS Elastic Kubernetes Service (EKS) is incredibly interesting. You can use EKS Hybrid to turn on-prem servers into Worker Nodes, but the Control Plane still exists within the cloud via AWS EKS. What’s amazing about this is the majority of compliance and security needs are around your data, and the data runs in Worker Nodes, which means you can still utilize EKS in the cloud at the Control Plane layer all while still staying compliant.
Closing Thoughts
Between the need for organizations to modernize applications in 2025 and performance/scalability concerns, hybrid cloud is the perfect intermediary solution to make it all possible. It’s a very exciting time to be in technology.
Top comments (0)