Skip to main content

Slinky: Bridging Slurm and Kubernetes

Welcome to Slinky, a set of powerful integration tools designed to bring Slurm capabilities into Kubernetes. Whether you’re managing high-performance computing (HPC) workloads or operating within cloud-native environments, Slinky helps bring together the best of both worlds for efficient resource management and scheduling.

What is Slinky?

Slinky is a set of projects developed by SchedMD to enable seamless interoperability between Slurm and Kubernetes. It introduces tools that allow users to run and manage Slurm clusters inside Kubernetes environments. This solution is ideal for hybrid compute scenarios, offering flexibility and ease of use for both HPC and cloud-native users.

Key Components:

    1. Slurm Operator
      The Slurm Operator is at the core of Slinky’s functionality. By utilizing aspects of Slurm, like job allocation/accounting/dependencies, fair-share, and priority scheduling, Slurm Operator successfully manages the scaling of Slurm nodes within Kubernetes.
    2. Slurm Bridge
      The Slurm Bridge is a future Slinky project, aimed to bring quick, intelligent scheduling of workloads across a Kubernetes cluster. This will allow for co-location of Slurm and Kubernetes workloads, bringing in the advantages of the Slurm scheduling and scale to both. Stay tuned for more information around this new Slinky project!

Why Choose Slinky?

Slinky provides significant benefits by integrating Slurm with Kubernetes, particularly for organizations managing both high-performance computing (HPC), cloud-native, and AI training workloads. Here are the key advantages:

Cloud capabilities icon

Seamless Hybrid Workload Management

Slinky enables the simultaneous management of HPC workloads using Slurm and containerized applications via Kubernetes on the same infrastructure. This is ideal for organizations running AI/ML training, AI/ML inference, scientific simulations, or data-intensive tasks alongside modern, cloud-native applications. It removes the need to maintain separate clusters, simplifying management and boosting efficiency.

High Throughput Computing - SchedMD

Dynamic Autoscaling

Slinky’s Slurm Operator allows automatic scaling of compute nodes. Slinky dynamically adds or removes Slurm nodes based on workload demands, ensuring that compute resources are efficiently allocated. This leads to reduced overhead costs and ensures that the right amount of infrastructure is available when needed​.

Prioritization fairness icon

Resource Optimization

Slurm Bridge brings the full capacity of Slurm’s node packing systems into Kubernetes. This allows for the intelligent scheduling of the right resources at the right time, leading to optimized resource allocation and better coordination between systems.

Slurm Installation Overview - SchedMD

Unified Infrastructure

Slinky makes it possible to run both Slurm and Kubernetes workloads on the same set of nodes, enabling better resource utilization. This is particularly useful in hybrid cloud environments where organizations aim to maximize the use of their infrastructure across various workload types without having to separate resources for each system

Placeholder icon1

Consistent Deployment

By using container images for Slurm components, Slinky ensures consistent and repeatable deployments across environments, whether on-premise, cloud, or multi-cloud. This allows organizations to package their applications and dependencies into containers, ensuring that they run identically in development, staging, and production.

Configure slurm environment icon

Flexibility & Customization

In Slinky, custom Slurm images play a crucial role in leveraging the power of Kubernetes while tailoring your Slurm environment to your specific needs. These containerized images allow users to configure Slurm with their unique dependencies, libraries, and configurations, providing consistency, flexibility, and control over compute environments.

Get Started with Slinky

Slinky is available to select SchedMD customers, with full commercial support. To learn more about the Slinky projects and Slinky support, reach out to SchedMD or visit the Slinky github page. 

Contact SchedMD SupportVisit Slinky Github

Recent Articles & Publications

September 20, 2024

What can a high-performance computer do?

August 1, 2024

Introducing the Slinky Project

March 26, 2024

Slurm releases move to a six-month cycle

Organize Your Workload Efficiently & Smoothly with SchedMD

Take your efficiency to the next level with Slurm from SchedMD. We can’t wait to do amazing things with you.

Request a Technical Call Today
Slurm Workload Manager - Download Slurm - SchedMd