Menu

Kubecon 2022: The Road Ahead Takes Kubernetes to the Edge

Applications at the so-called Far Edge often deliver a high return on investment, whether they’re driving more efficient oil extraction, virtual reality shopping or a more effective (and safer) factory floor.  But these deployments are complex and challenging to deploy and manage, encompassing thousands of distributed sites (and systems) with custom installations of hardware and software.

Kubernetes, with its native architecture of standardized code libraries and automated lifecycle management, is a natural tool for these locations.  This is particularly the case when lifecycle management tools scale and security tools harden to support and manage the thousands of sites covered by many of these applications.

But Kubernetes adoption at the edge has been limited.  The minimum resources required to operate a Kubernetes cluster or node can be too large for the power, resource footprint, network bandwidth or form factor of edge-based servers (or devices).  The inability to integrate unique devices or legacy systems makes deployment difficult.  Many customers cannot manage the customization required for the complexity of these deployments.

At the KubeCon 2022 in October, major Kubernetes solution providers made it clear that they intend to address these issues.  Two stand out the most, as Red Hat and SUSE introduced specialty configurations to serve the resource needs of the Far Edge.

Red Hat announced the developer preview (for early 2023) of Red Hat Device Edge. It’s based on MicroShift, a lightweight Kubernetes orchestration system intended for deploying and managing containers on small devices, including IoT and point of sale.  The company also released a version of Red Hat Enterprise Linux (RHEL) optimized for the edge, using minimum bandwidth for intelligent software updates.  Red Hat indicates that these offerings will allow for zero-touch provisioning, observability of overall system health and updates with automated rollback. Although the solution runs on bare metal to limit resource usage, it also allows for integration of legacy Windows-based applications using an embedded virtual machine.

SUSE made similar announcements but advanced the solution a step further.  To address resource requirements of the Far Edge, SUSE provides K3s, which it notes is the most widely adopted lightweight Kubernetes solution in the industry, and the company offers SLE Micro as a resource-efficient custom version of Linux.  K3s also supports a virtual machine for addressing legacy applications.  But, recognizing the complexity of creating and managing custom software stacks needed for the Edge, SUSE has also defined and sells a top-to-bottom layered stack of open-source software (including NeuVector security) to support the entire Far Edge infrastructure. That allows partners to easily plug in custom offerings to support unique industry requirements.  With an API-driven architecture (and the ability of Fleet to support lifecycle management of tens of thousands of nodes), partners and customers can manage the distributed environment with the automation tools necessary for success.

SUSE also announced future support of the open-source Project Akri.  Akri (originated by Microsoft and then shared with the CNCF community) enables automated discovery of downstream devices, from sensors to GPUs, for integration into a Kubernetes cluster.

These new edge-based solutions form the foundation of a toolkit that customers and partners can use to build and maintain Far Edge deployments.  But the remaining piece to complete the offering is expertise.  Both Red Hat and SUSE offer customer assistance with the proof-of-concept stage, but distributed solutions need support from POC through the multiple added stages of rollout (first 10 sites, next 100 sites, etc.) as well as ongoing management.  An expert force of partners who can deploy, maintain, and manage the thousands of custom systems is required. While oft-cited successful edge adopters at BMW, Lockheed Martin and other global enterprises have the resources to manage the unique investment and complexity of these environments, the more common commercial enterprise does not.  Widespread adoption will require assistance from Global System Integrators, industry specialists, and even local managed service provider to assemble the solutions and make them work.

Related Insights
Snowflake's SnowWork Targets the Gap Between Data Insight and Business Action
March 25, 2026

Snowflake’s SnowWork Targets the Gap Between Data Insight and Business Action

Brad Shimmin and Nick Patience explore Snowflake’s Project SnowWork and how the Agentic Enterprise Control Plane turns the AI Data Cloud into a "system of action" for autonomous workflows across...
Mistral Forge Takes Aim at RAG. But Who Actually Needs Custom Models
March 25, 2026

Mistral Forge Takes Aim at RAG. But Who Actually Needs Custom Models?

Nick Patience, AI Platforms Practice Lead at Futurum, examines Mistral Forge, a custom enterprise AI model training platform, and argues that while its approach is sound, the addressable market may...
Oracle Positions AI Database 26ai to Lead $1.2 Trillion Market by Bridging the Agentic Reasoning Gap
March 25, 2026

Oracle Positions AI Database 26ai to Lead $1.2 Trillion Market by Bridging the Agentic Reasoning Gap

Brad Shimmin and Keith Kirkpatrick of Futurum explore Oracle's pivot to agentic plumbing. Oracle is embedding autonomous reasoning directly into Oracle AI Database 26ai to solve the enterprise data latency...
Grounding the Agentic Mandate As the Semantic Layer Market Eyes 19% Growth, Microsoft Fabric IQ Targets Leaders Prioritizing AI Investment
March 20, 2026

Grounding the Agentic Mandate: As the Semantic Layer Market Eyes 19% Growth, Microsoft Fabric IQ Targets Leaders Prioritizing AI Investment

Brad Shimmin, VP and Practice Lead at Futurum, shares insights from FabCon and SQLCon 2026 on how Microsoft is leveraging the new Database Hub and Fabric IQ to unify transactional...
NVIDIA GTC 2026 Day 1 - Can NVIDIA’s Ecosystem Accelerate the Inference Inflection
March 18, 2026

NVIDIA GTC 2026 Day 1 – Can NVIDIA’s Ecosystem Accelerate the Inference Inflection?

Brendan Burke, Research Director at Futurum, breaks down NVIDIA GTC 2026 Day 1, highlighting the NVIDIA Vera Rubin platform, the $27B Nebius-Meta deal, and how partners like HPE and Micron...
NVIDIA Agent Toolkit
March 16, 2026

At GTC 2026, NVIDIA Stakes Its Claim on Autonomous Agent Infrastructure

Nick Patience and Mitch Ashley, analysts at Futurum, examine NVIDIA's Agent Toolkit announcements at GTC 2026, covering NemoClaw, AI-Q, the Nemotron Coalition, and what they mean for enterprise agentic AI...

Book a Demo

Newsletter Sign-up Form

Get important insights straight to your inbox, receive first looks at eBooks, exclusive event invitations, custom content, and more. We promise not to spam you or sell your name to anyone. You can always unsubscribe at any time.

All fields are required






Thank you, we received your request, a member of our team will be in contact with you.