The newly introduced image mode for Red Hat Enterprise Linux is a new deployment method that delivers the platform as a container image. Image mode takes a container-native approach to building, deploying and managing the operating system, providing a single workflow to manage the entirety of an IT landscape, from the applications to the underlying operating system, with the same tools and techniques. These updates are behind what Red Hat is calling reimagined Linux for the age of AI.
But before we dive into the new ‘reimagined’ Linux, let’s quickly remind ourselves how this operating system grew to become a solidified piece of terminology in and of itself. When you hear people saying ‘rel’ they’re actually saying ‘rhel’ with a silent h, or in fact RHEL as it should be thought of. Red Hat Enterprise Linux, or RHEL, is of course the company’s enterprise Linux platform composed of its operating system, tools, supporting functions and extensions.
But that time-honoured term may now need reinvention for the Artificial Intelligence (AI) and Machine Learning (ML) era. Might we now need to start saying AIRHEL? The company has used Red Hat Summit 2024 to showcase a new deployment option for the Linux platform that makes containers the language of the operating system, providing a single container-native workflow that spans both development and operations teams across the hybrid cloud.
For many organisations, the engineering team at Red Hat remind us that a Standard Operating Environment (SOE) or ‘gold image’ based on Red Hat Enterprise Linux forms the foundation of a firm’s technology strategies. These images power hybrid cloud deployments from the datacentre to public clouds across private on-premises clouds and out to the edge. As gleamingly functional as a gold image is, these deployments typically need to be customised to meet unique business and environmental needs.
What is a gold (golden) image?
In the music industry, a gold image is the final cut of an album or film after all edits and mixing have been completed. It’s in its final, perfect form, so it’s considered to be gold. According to Red Hat’s own topics webpages, this meaning has been carried over into systems administration.
“In this context, a golden image is an intentionally configured snapshot of a system, (server, virtual desktop environment, or even a disk drive) which can be used to deploy new instances. Because this golden image (or sometimes gold image) is used in network virtualisation to create new systems, it is also called a master image or clone image. Another popular term is a baseline image, which can be an illustrative term to frame why golden images are so useful: they create a consistent, reliable baseline for system configuration, which can make it easier to maintain those systems across their life cycle,” explains the company.
Moving beyond the allure of golden images then, the rise of AI workloads that require greater speed and flexibility mean that the operating system needs to respond by becoming more adaptable, more scalable and more responsive. To answer this, Red Hat explains that it is bringing the innovations powering modern application development practices and containers, to the heart of the operating system.
Also read: Red Hat optimizes OpenShift for hybrid AI and cloud
Containers in the ‘last mile’
Image mode for Red Hat Enterprise Linux builds on the success of open source projects such as bootc (a technology project devoted to act the key component in a broader mission of bootable containers) to create a deployment model that fits into container-native workflows. Image mode enables the operating system to use the same tools, skills and patterns as containerised applications, letting operations and infrastructure teams speak the same dialect as developers.
“The era of the AI-defined organisation requires us to reassess all technologies across the enterprise, not just those directly impacted by intelligent applications and workloads, said Gunnar Hellekson, vice president and general manager, Red Hat Enterprise Linux, Red Hat. “By delivering the world’s leading enterprise Linux platform in a container format, we’re providing an operating system that can match the speed, efficiency and innovation of the AI era, backed by the consistency and trust that CIOs have come to expect from Red Hat Enterprise Linux.”
Following the development of this story with some preexisting knowledge, we know that Linux already sits at the core of containers, but image mode takes Linux’s role a step further by making it possible to manage the entire operating system through container-based tooling and concepts, including GitOps and Continuous Integration/Continuous Delivery (CI/CD). This streamlined approach helps address challenges in managing Linux at scale, from pushing patches to disparate locations to disconnects between operations teams and the application development cycle.
Podman Desktop AI Lab
Nearly every organisation is planning for a future structured around AI workloads or, at the very least, applications infused with some level of AI capabilities. Red Hat is proposing image mode for Red Hat Enterprise Linux as a means of supporting the need to move more quickly when it comes to building, testing, and deploying AI applications, both through its flexible nature and its tight integration with Podman Desktop AI Lab.
According to developer resource repository GitHub, “The Podman Desktop AI Lab extension simplifies getting started and developing with AI in a local [computing] environment. It provides key open source technologies to start building on AI. A curated catalogue of so-called ‘recipes’ helps navigate the jungle of AI use cases and AI models.”
Red Hat details next here and notes that developers can build intelligent applications using Podman Desktop AI Lab on their laptop through a process that’s greatly simplified by the AI Lab’s recipe catalogue and straightforward AI playground environment. They can then use the bootc plug-in to readily convert to containers, bootable images or even bare-metal installers, all running on the backbone of Red Hat Enterprise Linux.
Additionally, Red Hat Insights now offers additional management features in support of image mode’s immutability. Operations teams can now view the deployment of operating system images across their infrastructure, while administrators are now able to update image mode systems directly from Red Hat Insights. In the future, image maintainers will also be able to better harden their images.
What to think about Red Hat
The resounding messages from Red Hat are pretty clear. The Red Hat that was acquired by IBM is still Red Hat i.e. the company is a somewhat IBM-style Red Hat in places (the penetration of IBM Watson AI at the coding tools level with IBM watsonx Code Assistant is undeniable and expansive), but it’s still Red Hat as we always knew it for the most part. Red Hat is also clear on its deployment of generative AI across its platform and tools (the Red Hat Ansible Lightspeed platform development is also significant) and the company has sent out a resounding message from the Red Hat OpenShift side to say that the cloud is hybrid, so, logically, so is AI. The subtle ‘we own this’ message is not lost on us guys, let’s continue to watch the evolution happening here.