Blog | Resource | Bigstack

CubeCOS: Cloud Native Virtualization Platform

Written by Brian Su | Oct 1, 2025 1:30:00 AM

Welcome to part one of the Getting to Know CubeCOS Series, where we explore CubeCOS, the open source platform, break down its components, and share practical guides to help you get started. In this blog post, I will cover CubeCOS, its origins, and what's next.

What is CubeCOS

CubeCOS is a free and open-source platform that brings together Infrastructure as a Service (IaaS) and Platform as a Service (PaaS). With opinionated best practices and pre-configured defaults, it enables you to set up production-ready infrastructure quickly. Built on OpenStack and tightly integrated with Kubernetes, CubeCOS automates your infrastructure stack to deliver a seamless, public cloud–like experience through its self-service portal.

Designed to scale from edge to data center, CubeCOS delivers reproducible environments that streamline resource management for developers and reduce operational overheads for administrators and operators.

Why we built CubeCOS: Industry shifts in cloud and virtualization

In 2016, we observed three trends in the cloud/ enterprise computing space that raised questions about the future of virtualized infrastructure.

Virtualization as the default

Baremetal was fading fast and replaced by the hypervisor, with the best options locked behind a proprietary stack.

Kubernetes scaling containers

As containerization matured, it was no longer a development or test tool. Kubernetes introduced orchestration to scale, while demanding a more flexible and robust underlying infrastructure.

Public cloud ecosystem maturing

Cloud providers stopped being "another choice for compute" and became an integrated ecosystem. But that didn't mean all workloads were suitable for the cloud.

Picture 1: Transitions on timeline of physical to virtual servers to container to orchestration to CubeCOS

These three shifts left us with a bigger question: is there a better solution that could tackle all three challenges? It became clear that the available solutions fell into three distinct camps, each with trade-offs that left a noticeable gap in the market.

The gap in open source virtualization platforms

From our experience, the on-premise virtualization infrastructure space can be divided into three categories:

  1. Proprietary enterprise solutions: Fully closed source solutions (e.g. VMware, Nutanix, or Microsoft), where the vendors retain tight control over their products and support. Often, with proven reliability, high costs, and low flexibility.
  2. Open source with professional service: Organizations deploy and manage open-source platforms (e.g., KVM, Kubernetes, or OpenStack) on their own infrastructure but rely on external experts or vendors for installation, customization, and ongoing support. The open source software itself is free, but professional services are paid for.
  3. Open source "as a service":  Vendors provide managed, cloud-like experiences built on open-source technologies. The vendor hosts or remotely manages the environment, abstracts the operational complexity, and charges a subscription fee for the managed service (e.g., Mirantis or OpenMetal). Here, customers consume open-source infrastructure in much the same way as a cloud service, rather than operating it themselves.

Each addressed part of the problem, but the gap remained: an out-of-the-box, reliable, open-source platform.

How we built our own open source cloud platform

The gap shaped our approach. We sought an open-source virtualization platform that offered an enterprise-grade experience without vendor lock-in.

Two development paths we considered

  1. Building CubeCOS on top of commercial solutions like VMware, leveraging their foundation and focusing CubeCOS on orchestration.
  2. Building CubeCOS from the ground up.

Picture 2: Choosing between integration on commercial foundations vs building CubeCOS from scratch

The first path risks locking us into someone else's product and the platform's constraints, which was what we wanted to avoid. We chose the second path to take complete control, selecting, integrating, and operating every component ourselves.

With that decided, we started designing our requirements: open source, production-ready, compatible with modern workloads, and maintainable. We mapped capabilities to existing projects, including hypervisor, storage backend, networking, orchestration, and monitoring, and filled any gaps with additional tooling.

Then came the implementation phase, which involved deploying, operating, and maintaining our platform. This is where we hit the first challenges.

Challenges in building with open source projects

Our challenges can be divided into three major categories that we aim to tackle with CubeCOS.

1. The Challenge of integrating open source components

To meet the defined requirements, CubeCOS required over 40 open source integrations. Integrating the individual tools and projects manually was arduous, time-consuming, and prone to human error.

2. Operating and maintaining

Working with separate command-line operations, management interfaces, and domain knowledge taxed our developers and operators.

3. Managing lifecycle

We didn't have a reliable way to test and validate new features or upgrades without risking breakage. This limited the long-term usability and maintainability of CubeCOS.

Picture 3: Illustration of orchestrating all the relevant components required

We could deploy the stack, but the challenge was the ongoing integration and maintenance of the platform. The available tools we explored solved deployment or integration, but rarely both. Most of the tools we tried solved one piece of the puzzle, but none could manage the entire lifecycle.

Hex Bedrock: Orchestration and automation for CubeCOS

That's why we built Hex Bedrock, a custom framework that brings reproducible deployments, service orchestration, and operation automation. We will cover how it works in our upcoming CubeCOS Deep Dive Series.

At a high level, Hex is the core foundation of the CubeCOS platform, providing:

  1. Integration modules: Connect and configure all required CubeCOS components.
  2. Unified management CLI: One interface for all cluster operations, replacing multiple disparate tools.
  3. Service health probes: Monitor services and perform repairs on unhealthy services.
  4. Reproducible cluster framework: Build consistent environments using pre-defined policies.

Subscribe to our blog to be notified when the upcoming deep dive series explores how Hex integrates OpenStack services.

With Hex, CubeCOS becomes easier to deploy, operate, and maintain at scale. Next, we’ll look at the problems CubeCOS addresses and the types of workloads it supports.

Challenges CubeCOS solves in deploying Kubernetes on OpenStack

CubeCOS targets one primary problem: the operational complexity of OpenStack. By abstracting away the deployment, integration, and lifecycle management, it delivers a consistent experience for administrators, operators, and developers — whether the workload is small-scale or cloud-scale.

Simplifying OpenStack infrastructure

Through Hex, CubeCOS handles the integration, orchestration, and operation of the OpenStack cluster. Delivering an operational, production-ready environment without manual service setup or much configuration.

Running Kubernetes on OpenStack

CubeCOS integrates Kubernetes natively, allowing clusters to consume OpenStack resources directly. Each Kubernetes cluster can be placed in individual tenants, with isolated compute, storage, and network allocations, to maximize utilization and support multi-tenant, cloud-scale workloads.

Avoiding vendor lock-in

CubeCOS is 100% open source under the Apache 2.0 license. It is designed so that the community building it can run, maintain, and extend it.

By addressing deployment complexity, Kubernetes integration, and vendor lock-in, CubeCOS will power your virtualization infrastructure. We have proven CubeCOS in production across different environments and scales.

Where can you deploy CubeCOS?

Over the past few years, we have deployed CubeCOS as a closed-source platform, using production feedback to refine and align it with enterprise requirements. Our partners and adopting organizations have deployed CubeCOS for:

Private cloud

Act as the infrastructure foundation for both virtualized and containerized workloads.

Public cloud

Serves as the foundation platform for providers delivering multi-tenant IaaS and PaaS offerings.

Edge deployments

Optimized builds that deliver the same capabilities in a reduced footprint for location-specific workloads.

Tell us how you apply CubeCOS. Share your workloads and use cases with us on GitHub Discussions or our community channels!

What can you expect in the future for CubeCOS?

We are committed to following the Four Open principles (open source, open design, open development, open community) defined by the OpenInfra foundation and to growing the CubeCOS community into a self-sustaining, community-led project.

In the upcoming releases, our road map includes:

  1. Improved platform stability: With expanded health check automation.
  2. Improved UX: Migrating CLI operations into the web UI for faster adoption and reduced learning curve.
  3. Storage flexibility: Support for external SAN-backed storage systems.
  4. Enhanced monitoring: Providing customizable thresholds and parameters for alerting to improve monitoring capabilities.

The CubeCOS community is growing, and your participation matters. Contribute, report bugs, share ideas, and help shape the future of open on-premise cloud infrastructure.

Join the CubeCOS open source community

Collaborate with us on GitHub discussions, view the list of open issues, and join our community channels.

Frequently Asked Questions

How is CubeCOS different from other cloud virtualization platforms?

CubeCOS is a free and open-source cloud virtualization platform that simplifies OpenStack deployment and provides native Kubernetes integration. It offers a cloud-like experience for running virtual machines and containers across edge and private cloud environments.

What challenges does CubeCOS address in the virtualization platforms space?

+

CubeCOS addresses the complexity, fragmentation, and vendor lock-in typically found in traditional virtualization platforms. It fills the gap with a fully open source, enterprise-ready stack built on OpenStack, designed for simplicity and flexibility.

How does CubeCOS simplify OpenStack operations?

+

CubeCOS uses Hex Bedrock, a custom orchestration framework that automates deployment, integration, and operation. It reduces operational overhead and enables reproducible lifecycle management at any scale.

Where can I deploy CubeCOS?

+

CubeCOS can be deployed in private clouds, public clouds, and edge environments. Designed to provide flexibility and power for both large-scale data center deployments and resource-constrained edge environments.

What is next on the CubeCOS development roadmap?

+

Upcoming features include support for external storage systems, deeper health automation, a more intuitive UI experience, and advanced monitoring capabilities in future releases. CubeCOS is developed in the open with a growing, community-led ecosystem.