
AI Networking: What GPU Infrastructure Really Requires
Learn how AI workloads change network design across GPU clusters, storage, east-west traffic, RDMA/RoCE, Kubernetes placement, and production inference.
Read field noteField notes / Latest
Engineering notes from operating open infrastructure: the failures, design decisions, and upstream work that make open infrastructure better.
Browse all field notes
Learn how AI workloads change network design across GPU clusters, storage, east-west traffic, RDMA/RoCE, Kubernetes placement, and production inference.
Read field note
Booting a GPU server isn't the same as making it production-ready. How Atmosphere unifies OpenStack, Ironic, and Kubernetes to turn H200 hardware into recoverable, reusable AI infrastructure.
Read field note
Compare GPU VMs, bare metal, Kubernetes, and managed inference for AI workloads. Learn which architecture fits your performance, control, scaling, and operational needs.
Read field noteTrends, best practices, and technical deep dives on open source cloud infrastructure.

Managing an on-premise cloud is complex, requiring time and resources for deployment, security, and scalability, often diverting teams from their core goals.

Questions linger around essential features such as high availability for VMs, entering maintenance mode, the efficiency of DRS, the nuances of storage integration, and among others, the simplicity of updates.
CPUs can’t keep up with AI workloads—but GPUs can. See how OpenStack, Kubernetes, and PCI passthrough supercharge AI/ML performance while ensuring seamless deployment and scalability.
The latest release of Atmosphere, now supporting the OpenStack Dalmatian, features a suite of enhancements. Key updates include a new streamlined upgrade process, advanced Keystone role configurations for Keycloak realms, and enhanced Neutron policy checks for address pair management. Additionally, significant security upgrades have been implemented for the Horizon service, alongside performance improvements in networking with OVN and DPDK interface configurations.
As businesses seek greater flexibility, cost efficiency, and vendor independence, many are migrating from VMware to OpenStack. However, transitioning between cloud platforms can be complex, requiring careful planning, seamless integration, and minimal downtime. That’s where Atmosphere comes in.
In today’s digital landscape, agility and efficiency are paramount for businesses of all sizes. Traditional network architectures, often rigid and complex, can struggle to keep pace with the ever-changing demands of modern applications and cloud environments.
Boost your cloud performance with Atmosphere's latest releases: v1.13.11 for OpenStack Zed, v2.2.11 for Antelope, v3.2.12 for Bobcat, and v4.2.12 for Caracal. These updates bring seamless operations, improved resource management, and fortified security to your cloud environment.
Running Kubernetes at scale isn’t always straightforward. Atmosphere makes it easier with better networking, storage, and automation. See how it simplifies management and scaling.
Cloud providers come in all sizes, and it’s easy to assume that bigger is always better (hint: it isn't!). In this blog, we’ll be taking a look at some of the debunking some of the most common myths about smaller cloud providers.
These updates and fixes are part of the ongoing maintenance and improvement of Atmosphere ensuring that it remains reliable, compatible with other tools and systems, and user-friendly for developers who work with the platform's infrastructure.