Choosing your environment
Local
Run RAFT locally when:
you want to run directly from a workstation, server, or compute cluster
you want the simplest installation and debugging loop
RAFT can run locally whether you are running on a single machine or on a cluster
with a job scheduler such as SLURM. Users should ensure an appropriate profile
for their compute environment is available in their ~/.nextflow/config
file. RAFT provides a minimal default profile upon installation.
Note
Users can specify which Nextflow profile to use by passing the profile
identifier with --profile to raft run (e.g. raft run ... --profile
my_favorite_profile).
For information on running RAFT on local hardware, see Running RAFT locally.
Hybrid
Choose this path when:
you want to launch RAFT from a local installation while offloading compute to the cloud
you want to run on AWS
In this mode, you install and run RAFT locally but pass --cloud to
raft run to execute the pipeline processes on cloud infrastructure. This
gives you the convenience of a local CLI while leveraging cloud compute
resources for scaling.
Note
Hybrid mode is the recommended way to run RAFT on AWS.
Cloud
Choose this path when:
you want to run on your own GCP infrastructure
you do not want to install RAFT yourself
In this mode, users set up RAFT projects with the BYOC Cloud Launcher and a RAFT runner image.
For information on running RAFT on GCP, see Running RAFT on the cloud.