Cloud, Cloud Native or What?
Sometimes it’s not easy to keep track of all the terminology in the IT world, but let’s try to properly classify the most important cloud-related terms. I should mention that even today, some experts mix things up, and the separation between terms isn’t always as clear as one might wish—but more on that later.
The good old on-premises installation
The term ‘on-premises’ generally means that an application is installed locally at the customer’s site. This might sound familiar because, until the late 2010s, this was the standard. To describe it concretely: in an on-premises setup, the servers and applications physically reside at the customer’s location along with their data. The customer can access their hardware at any time, which can be an advantage—but not necessarily—if they need assistance.
And then came the cloud
There are countless interpretations of the term ‘cloud.’ The word itself suggests that certain services or applications are provided somewhere, somehow. How this is implemented behind the scenes usually doesn’t concern most users, as long as certain quality standards and rules are maintained. But of course, we’re interested—so let’s dive into the topic with a few questions.
Is an internal company data center a cloud?
Yes and no. Sometimes you can see it that way, and sometimes that’s correct – but not always. If only internal applications are operated without scalability, virtualization, and automation, then it’s more like a conventional server-client setup where the server resides in the data center, and the client runs locally. However, if the data center provides cloud services with the characteristics of a cloud, then it can indeed be called a private cloud.
What are cloud services and the characteristics of a cloud?
Let’s start with the glorious three (types of cloud services):
Infrastructure as a Service: This means, ‘Give me as many hardware resources as I’m willing to pay for through virtualization, and I’ll handle the rest myself.’ Bonus points if you noticed the word ‘virtualization.’ Of course, these aren’t real hardware resources—you’re allocated a portion of actual hardware (processor, memory, SSD, network, etc.). Often, you can choose the location or data center where the physical machines reside, but special requests naturally cost extra.
Platform as a Service: Here, things are already more convenient for the customer. You select the applications you want to run in the cloud, and the cloud provider takes care of everything else—hardware infrastructure, operating system, and updates.
Software as a Service: For the user, it should feel as if they’re launching their favorite app locally after installation. The only thing they might need to configure is the in-app settings—everything else is handled by the cloud provider.
What else is there beside the private cloud?
The private cloud provides cloud services in a secure data center only to authorized individuals and groups. It doesn’t necessarily have to be one or more company-owned data centers; governments and government organizations also operate private clouds in practice. Naturally, these have special requirements for security and data protection.
The public cloud also has strict security requirements and rules. However, unlike private clouds, it’s available to everyone. For a fee, anyone can secure the services they want from the ‘Big Three’—AWS, Azure, and Google Cloud—or others for a limited time. Ultimately, you pay for the use of virtualized hardware and services over a certain period. The longer and more resources you use, the more expensive it becomes—but there are discounts for long-term commitments, just like any other subscription model.
The hybrid cloud—yes, you guessed it—is a mix of public and private clouds. Sometimes it’s desirable to outsource certain services even if a company operates its own data centers, or to keep certain data in-house. For such cases, the hybrid cloud concept offers the best of both worlds.
What are the characteristics of a cloud, and what was that about cloud native?
The typical characteristics of a cloud are:
Scalability: This means that when many users want to use an application, cloud technologies ensure that additional instances are launched as needed to handle the load. Conversely, it can scale down—even to ‘scale to zero,’ meaning no instance is running at all. That’s more of a ‘serverless’ concept, which we’ll explain another time.
Virtualization: First, a quick detour into containers. An image—the blueprint for a container—is defined by a file, usually a Dockerfile. Details like layers, Linux namespaces, and cgroups we’ll skip for now, but suffice it to say Linux and Docker achieved something revolutionary. With a container runtime, you can run applications inside a container in exactly the same isolated environment every time. Ultimately, this means virtualized resources in the cloud can be further divided on virtual machines in a resource-efficient way. Nice, isn’t it?
Automation: If I can scale such containers automatically during high workloads or restart them in case of failure—known as resilience—without user interaction, then a high degree of automation is achieved, which is one of the core features of a cloud solution.
What is cloud-native?
Finally, the term ‘cloud native’ encompasses technologies and architectural concepts (such as microservices) aimed at optimally supporting the cloud characteristics mentioned earlier and others not explicitly listed. For example, you can run applications using Kubernetes—the leading cloud-native technology according to CNCF—even if the applications themselves don’t follow the cloud-native paradigm. That’s why it’s sometimes hard to distinguish between cloud and cloud native.
I hope this gave you a clearer picture of what ‘the cloud’ means in IT. Stay tuned!