1. Object-store for NAME

All tasks are the responsibility of CEDA except for those that are prefixed by "MO - " (Met Office):

[x] Agree details: label, size=100TB, etc
[x] Request set-up
[ ] Agree naming conventions for: buckets, objects
[x] Provide scripts/code for populating the object-store
[ ] MO - manage population of the object-store
- Possibly create an ERA5 archive for NAME
[ ] Investigate how Fortran might communicate/retrieve data from the object-store (default: GET and cache locally)
[ ] Investigate async GETs for faster retrieval
[x] Provide scripts/code for retrieving files from the object-store
[ ] Document the interface, code and the process

2. NAME Web Processing Service

We will use the previously developed NAME WPS: swallow.

All tasks are the responsibility of CEDA except for those that are prefixed by "MO - " (Met Office):

[x] Review the existing tools/software: CEDA WPS User Interface, swallow WPS processes
[x] Agree required changes to: CEDA WPS UI, swallow
[x] MO - provide compiled NAME code and any other required files/resources
[x] NAME will be configured to run on a single task/process (but is compiled with MPI by default since NAME v8), and can use parallelism using OpenMPI
- [x] Update Conda env to ensure that MPI libraries are available
[x] Create test runs for NAME within process unit tests
[x] Update swallow with agreed changes, and update tests
[x] Update CEDA WPS UI to provide automated web-form for swallow jobs
[x] Update the Ansible playbook(s) for deploying
[x] Deploy prototype on staging server
[x] MO - review prototype
[ ] Respond to prototype review
[ ] Document the service (internally/externally):
- [ ] CEDA Services site
- [ ] Create a Help Page on CEDA/JASMIN help site
[ ] Agree maintenance and development plan
[ ] Provide instructions/method for MO to be able to develop/test the service
[ ] Deploy first release on production cluster