High throughput and high performance computing without servers or clusters
Run containerised jobs across many clouds without managing infrastructure
Run jobs from anywhere
Manage jobs using a simple batch system style command line interface or interact programmatically using a RESTful API.
Invisible infrastructure
All infrastructure provisioning is handled completely automatically and is totally invisible to the user.
Reliability & reproducibility
All jobs are run in containers to ensure they will run reliably anywhere and are reproducible.
Workflows
Construct complex workflows by specifying the dependencies between different jobs.
MPI ready
Run multi-node Open MPI, Intel MPI or MPICH jobs in addition to single-node jobs.
Data access
Jobs can access data transparently from OneData, WebDAV or object storage.
A serverless experience for HTC & HPC
Just run your jobs, there is no need to create or manage clusters
Multi-cloud native
Go beyond a single cloud and leverage the resources and features available across many clouds simultaneously. No VPNs, tunnels or any special networking is required.
One cloud or many
The most appropriate cloud for each job is used, based on user-specified requirements and preferences.
Improve resource utilisation
Run jobs on across multiple clouds opportunistically, utilising idle resources
No more login nodes
Access any computing resources, from anywhere