Professional services Managed service SaaS

nAG Storm

Observability

Cluster & queue metrics
Job & usage analytics
Dashboards & alerts
Audit trail

Budget control

Per-user & project budgets
Showback per job
Scale to zero & spot tiers
Idle shutdown

User experiencenAG

Browser desktop GPU visualization File manager Web portal Guided onboarding

ApplicationsnAG

Spack Intel oneAPI NVIDIA CUDA Lmod modules Containers nAG Library

SlinkyOpen source · SchedMD

Slurm, running natively on Kubernetes

slurmctldscheduler

slurmdbdaccounting

slurmrestdREST API

NodeSetsCPU · GPU · Spot

LoginSetslogin & desktop nodes

Autoscalingnodes follow the queue

KubernetesAny CNCF cluster

Any cloud, or your data centrenAG

Google Cloud AWS Azure On-premises NVIDIA GPUs

Data fabric

Shared POSIX storage
Parallel filesystems
Object storage
Hybrid & cross-cloud
Global namespace

Landing zone

Your cloud org & policies
Single sign-on & identity
Egress control & WAF
Secrets & your keys
GitOps: all as code

Slinky at the core. nAG everywhere around it.

nAG Storm

HPC as a platform,
built on Slinky

The scheduling HPC trusts, with Slurm on Kubernetes at the core — and the engineering nAG adds to make it a platform your users, finance team and security team can all live with.

Press → or tap to begin

The core · open source by SchedMD

Slinky: Slurm, native to Kubernetes

Slurm's scheduling, running as Kubernetes workloads. Users keep the tools they know; operators get Kubernetes' elasticity and automation.

slurmctld, slurmdbd and slurmrestd run as managed Kubernetes workloads.
NodeSets define compute: CPU, NVIDIA GPU and Spot pools.
LoginSets provide login and interactive desktop nodes.
The same sbatch, srun and sacct users already know.

nAG adds · portability

Runs anywhere Kubernetes runs

One platform definition, deployed on the substrate that suits each customer — public cloud, sovereign cloud or their own data centre.

Google Cloud, AWS, Azure or on-premises Kubernetes.
Elastic node pools that scale to zero: compute costs money only while jobs run.
NVIDIA GPU pools for AI and visualization, started on demand.
Spot / preemptible tiers for cost-efficient throughput.

nAG adds · data

Data where the work is

Compute is elastic; data has gravity. nAG integrates the storage a customer already has and moves data to where it is needed.

Shared POSIX home and project storage, the same on every node.
Parallel filesystem and object storage integrations.
Hybrid and cross-cloud data movement.
A global data fabric: one namespace across sites and clouds.

nAG adds · landing zones

Built for enterprise landing zones

Deployed into the customer's own cloud organisation and network, under their policies — not a black box beside them.

Respects the organisation's policies, networks and customer-managed encryption keys.
Single sign-on with the customer's identity provider; POSIX identity handled automatically.
Egress control, web application firewall and secrets management.
Everything as code with GitOps: auditable and repeatable.

nAG adds · applications

Applications, ready to run

A scheduler is only as useful as the software on it. nAG brings decades of numerical and HPC application expertise.

Curated software stacks built with Spack, served as environment modules.
Intel oneAPI and NVIDIA CUDA toolchains; containers where they fit.
The nAG Library and nAG's numerical expertise.
Application onboarding and performance engineering by nAG HPC specialists.

nAG adds · user experience

A desktop for every user

From sign-in to first job in minutes, from any browser, with nothing to install.

A full Linux desktop in the browser, one per user, started on demand and stopped when idle.
GPU-accelerated visualization for pre- and post-processing.
A web file manager and a self-service portal.
A guided tutorial that walks new users through their first Slurm job.

nAG adds · cost control

Budget control, built in

Elastic compute without surprise invoices — limits enforced by the scheduler, visibility for finance.

Per-user and per-project compute budgets, enforced by Slurm.
Showback per job, per user and per project.
Scale to zero, spot tiers and idle shutdown by default.
Cost dashboards for finance as well as engineers.

nAG adds · observability

Observability, end to end

Know what is running, for whom, and what it costs — from the queue down to the node.

Cluster, queue and job metrics in shared dashboards.
Usage analytics and alerting.
An audit trail of who ran what, where and when.
Platform health monitored by nAG.

nAG adds · delivery

Delivered the way you want it

The same platform, three ways to have it.

Professional services — nAG builds Storm in your cloud or data centre, and your team runs it.
Managed service — nAG operates Storm for you, in your environment.
SaaS — Storm as a service: sign in and run.

Scan to try it liveFree 4-hour trial: a Slurm desktop in your browser.storm.nag.com