CIQ Fuzzball integration with PBS Professional and Slurm bridges AI-era infrastructure with traditional HPC clusters

New provisioner allows enterprises to leverage Fuzzball's container orchestration while preserving existing scheduler workflows and operational investments

RENO, Nev., – December 4, 2025 – CIQ, the founding support and services partner of Rocky Linux and a leader in high-performance software infrastructure, today announced that Fuzzball, its container orchestration platform for performance-intensive computing, now integrates with PBS Professional and Slurm workload managers as backend provisioners. This integration enables organizations running traditional HPC clusters to adopt Fuzzball's modern container orchestration and workflow capabilities while maintaining continuity with existing scheduler interfaces, or have the option to leverage Fuzzball's intuitive web-based workflows when ready.

For decades, PBS and Slurm have been the backbone of HPC computing at research institutions, national laboratories and enterprises worldwide. However, as organizations face increasing pressure to support both traditional simulation workloads and emerging AI/ML applications, they're caught between the need to modernize and the risk of disrupting mission-critical operations. Fuzzball's provisioner integration with PBS and Slurm solves this fundamental challenge by allowing these established schedulers to provision resources for Fuzzball workloads, creating a bridge to container-native infrastructure without requiring wholesale replacement of proven systems.

"HPC centers have invested years building operational expertise, user training and workflow automation around their existing schedulers," said David Godlove, senior HPC engineer at CIQ. "This integration gives organizations the best of both worlds. Users can continue submitting jobs directly to Slurm or PBS exactly as they always have, while gaining access to Fuzzball's modern workflow catalog, container orchestration and hybrid cloud capabilities. It's about choice and incremental modernization, not forced change."

The integration leverages Fuzzball's new provisioner configuration system, which now allows administrators to designate PBS or Slurm clusters as resource providers alongside static compute pools and cloud resources like AWS. This unified approach delivers three key benefits for HPC organizations:

• Reduce adoption risk for enterprise decision-makers: IT leaders can adopt Fuzzball without requiring budget for staff retraining, workflow rewrites or infrastructure replacement. Users familiar with PBS or Slurm commands can continue using them while gradually exploring Fuzzball's workflow capabilities. The ability to leverage existing scheduler investments dramatically lowers the financial and operational risk of modernization initiatives.
• Support hybrid workload requirements: As AI and machine learning workloads increasingly run alongside traditional HPC simulations, organizations need infrastructure that bridges both paradigms. Fuzzball's provisioner integration allows teams to execute containerized AI workflows while traditional users continue their established job submission patterns. Administrators gain unified visibility and control across all workloads through Fuzzball's interface, regardless of how jobs are submitted.
• Enable fine-grained resource control: The integration includes powerful policy expression capabilities that give administrators precise control over resource allocation based on user identity, job requirements, organizational units and workload characteristics. Administrators can fully define resource specifications including CPU and GPU counts, memory requirements and cost per hour, then use policy expressions to route workloads intelligently across heterogeneous infrastructure including legacy clusters, cloud resources and specialized hardware.

"Organizations shouldn't have to choose between innovation and stability," added David Godlove. "This provisioner integration is about giving HPC centers a practical path forward, one that respects their existing investments and operational workflows while opening the door to container orchestration, modern workflow management, hybrid cloud capabilities and the infrastructure flexibility that both traditional research and emerging AI workloads demand."

With the PBS and Slurm provisioner integration with Fuzzball, organizations can configure their existing schedulers as backend resource providers while adding Fuzzball's workflow catalog, drag-and-drop workflow editor, automated data management and unified job monitoring to their computing environment, all without disrupting current user workflows.

For organizations interested in learning more about Fuzzball's integration with PBS Professional and Slurm, as well as additional improvements to Fuzzball over the past year, visit the CIQ blog, ciq.com/products/fuzzball or contact info@ciq.com.

About CIQ

CIQ delivers secure and performant software infrastructure for the demands of all modern workloads, from the most mundane to the most extreme HPC and AI jobs. We believe infrastructure should drive the future of your business and that both the operating system of a single machine and the orchestration layer to manage a cluster of machines and even hybrid environments needs to be optimized for your requirements. We are an open source company who has started and/or contributed to critical infrastructure projects such as Rocky Linux, Warewulf, Fuzzball, Ascender and Apptainer. For more information, visit ciq.com.