What is Azure CycleCloud?
Azure CycleCloud is a tool for orchestrating and managing High Performance Computing (HPC) environments on Azure. It provisions the required infrastructure, deploys familiar HPC schedulers and automatically scales clusters so jobs run efficiently at any scale. With CycleCloud you can create different file systems and mount them to the compute nodes to support HPC workloads.
Organizations with long-standing HPC environments often build deep expertise and custom tooling around a specific scheduler. CycleCloud solves the problem of moving these setups to Azure without a redesign: it abstracts the Azure building blocks such as VMs, scale sets, network interfaces and disks, so HPC administrators keep working with the cluster and scheduler they already know. CycleCloud is not a scheduler itself but a management platform for deploying and operating HPC resources.
Core Features
- Scheduler-agnostic orchestration: Built-in support with autoscaling plugins for Slurm, PBS Pro, IBM LSF, Grid Engine and HTCondor. You can integrate your own scheduler through the autoscaling REST API.
- Automatic scaling: CycleCloud adjusts cluster size and components based on job load, availability and time requirements, so users do not need to build their own autoscaling routines.
- Declarative templates: Complete HPC environments including NFS servers, parallel file systems, authentication and license servers are defined through a single management plane and can be shared and replicated.
- Azure integration and APIs: Integrates with Azure Monitor and Microsoft Cost Management as well as services such as Azure NetApp Files and Microsoft Entra ID, complemented by a RESTful API and a built-in CLI.
Typical Use Cases
Migrating existing HPC environments: Teams move long-standing on-premises setups with established schedulers to Azure without rebuilding their workflows and tools.
Simulation and research: Scientific and engineering simulations run on clusters that grow on demand and shrink again once the jobs complete.
Reproducible environments for multiple teams: HPC setups defined as templates are rolled out consistently across different teams and projects.
Benefits
- Keep using familiar schedulers instead of learning new tooling
- Automatic scaling reduces idle costs and manual effort
- No license fee for CycleCloud, billing only for the Azure resources you use
- EU regions available for workloads with data sovereignty requirements
Integration with innFactory
As a Microsoft Solutions Partner, innFactory supports you with the adoption and operation of this service.
Typical Use Cases
Frequently Asked Questions
What is Azure CycleCloud?
Azure CycleCloud is a tool for orchestrating and managing High Performance Computing environments on Azure. It provisions the infrastructure, deploys familiar HPC schedulers and autoscales clusters based on job load. CycleCloud is not a scheduler itself but a platform for building and running complete HPC systems.
When should I use Azure CycleCloud?
CycleCloud fits when you move an existing HPC environment with an established scheduler such as Slurm, PBS Pro, LSF, Grid Engine or HTCondor to Azure while keeping your familiar workflows. It also suits simulation and research workloads that need clusters scaling on demand, as well as reproducible HPC setups shared across multiple teams.
How much does Azure CycleCloud cost?
There is no license or subscription fee for CycleCloud itself. You pay only for the underlying Azure resources such as VMs, scale sets, disks, storage and networking on a pay-as-you-go basis. Spot VMs and reserved capacity can be used to further reduce compute costs.
Which HPC schedulers does Azure CycleCloud support?
CycleCloud is scheduler-agnostic and provides built-in support with autoscaling plugins for Slurm, PBS Pro, IBM LSF, Grid Engine and HTCondor. You can also integrate your own scheduler through the autoscaling REST API.
