Cloud architect & SRE. Former Scaleway CTO, current Brainboard CEO.
It is fascinating how cloud infrastructure has evolved and how we as an industry shifted (processes and mindsets) from hardware into software, as well as how we were able to build stable and performant components that are, in many cases, better than the hardware.
We even replaced our implicit use and meaning of the word “infrastructure.” When we use it, we often refer to cloud infrastructure only.
It came as no surprise that this occurred due to our utilization of software, which we can control more effectively. Additionally, software offers a shorter testing and release cycle compared to hardware, with the added benefit of being potentially cheaper (at least in theory).
So, is it all software now?
To better understand the big picture, we need to look at the parts of the iceberg.
• The Tip: Users and consumers interact mainly with APIs and virtual resources for computing, networking, storage and security. We’re increasingly moving away from hardware to the extent that some might consider software the predominant aspect of cloud computing.
• The Body: This includes all the underlying hardware that is abstracted from the users and has been virtualized (with software).
For cloud providers, everything is the same as before. They still have data centers, racks, cables, fibers, physical servers, disks and so on.
Now, with the rise of AI and the demand for huge computing power, hardware is back in the ring with more weight as GPUs are becoming the new “cloud” for AI.
Commoditization is inevitable.
Now, the picture is clear: Software is overtaking hardware, and the cloud enables the consumption of resources as commodities, easily abstracted through APIs. Provisioning and destroying resources now takes a few seconds with one simple API call.
It’s all about abstractions.
Although it may seem new, this abstraction is actually a fundamental aspect of infrastructure and software evolution. It began with the operating system, which abstracted hardware, then progressed to virtual machines abstracting the operating system, followed by containers abstracting virtual machines, and now culminates in cloud resources, abstracting every aspect of infrastructure.
Does it end here?
These abstractions emerged as answers to and natural progressions from our limitations. Every time we create a new technology, we abstract it to enhance its capabilities once it reaches a certain level of maturity.
This pattern has persisted since assembly code, through languages, operating systems, clusters and now cloud services.
So far, I’ve talked about the historical evolution of infrastructure. Let’s focus now on what lies ahead, exploring the forthcoming (re)evolution of cloud infrastructure.
Cloud-native applications or app-native infrastructure?
The most important use of cloud infrastructure is to enable applications to run effectively and developers to build with the most secure, cost-effective and easy-to-manage stacks while reducing the time to market.
To further narrow the gap between development and cloud infrastructure, the industry introduced what’s known as cloud-native applications. In this approach, the application’s architecture aligns closely with that of the infrastructure, with about 99% similarity. For this, most people utilize off-the-shelf components like databases, API gateways, storage, security and load balancers, eliminating the need to reinvent the wheel.
Put another way, we may think that the application is adapting to the cloud infrastructure, but when we look at it closely, the infrastructure is actually serving the application and not the other way around.
We in the industry construct cloud infrastructures that closely align with the application and support its architecture, leading to the term “application-native infrastructure.” This term accurately describes the evolution where we’ve transformed our approach to building alongside the infrastructure.
This shift, unlocked by APIs, has created a new mindset and introduced tools where the infrastructure is essentially defined by a YAML manifest detailing its target states.
Then why YAML and infrastructure as code tools?
The industry shifted its mindset but not its methods. We transitioned to a fully API and software-based infrastructure. However, many individuals who are used to building and managing this infrastructure lack a developer background. They needed something more accessible to learn, understand and use. That’s why YAML-based configuration tools became and remain popular.
This mindset and these tools represent a new abstraction of the API, where we describe the desired state in almost plain English text (in a certain format), and the tools handle the necessary API calls to make it happen.
However, this comes with a steep learning curve for engineers and complexities in managing environments. Once again, we find ourselves reaching the limits of our current model, and another evolution is taking place.
Human-friendly abstractions are the future.
With the rise of AI, we’re abstracting complexity by embracing technologies that resonate with human intuition. Take ChatGPT, for instance. We can simply articulate our goals in plain English, and it generates code for provisioning the infrastructure accordingly.
Another approach is using visualization. For example, with Brainboard, you can draw your cloud infrastructure, and the necessary deployment and management code is automatically generated.
These examples illustrate the next-generation software and mindset. The shift is happening now, and the next set of tools will be adapted and optimized for humans.
We’re moving away from requiring humans to learn coding languages, infrastructure APIs and YAML files, and toward technology adapting to our requirements, generating the intricate details necessary to fulfill our requests.
This new abstraction will give birth to a completely different approach to interacting with cloud infrastructure, potentially even challenging the necessity of current low-level cloud components.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
