Standardizing names for storage and cloud environments.

What is inaccurate?

We currently use many terms for individual and shared environments on Cheaha. Birhanu and I put together the following:

- lab space:                 GPFS shared
- personal account:          probably LTS individual
- lab account:               probably LTS shared
- cheaha project space:      GPFS shared
- project space:             GPFS shared
- project storage:           GPFS shared
- project shared space:      GPFS shared
- user space:                probably GPFS individual
- lts space:                 either LTS individual or LTS shared
- lts lab account:           LTS shared
- Lts Storage space:         either LTS individual or LTS shared
- personal lts account:      LTS individual
- lab lts account:           LTS shared
- lts Shared Space:          LTS shared
- Cheaha Storage:            GPFS shared
- Project Directory:         GPFS shared
- LTS storage:               either LTS individual or LTS shared
- LTS account:               either LTS individual or LTS shared
- LTS Lab allocation:        LTS Shared
- LTS Shared account:        LTS Shared
- Individual Storage:        LTS individual
- Shared storage:            probably LTS Shared
- Shared LTS account:        LTS Shared
- Shared Lab Space:          GPFS shared
- Shared Project Storage:    GPFS shared

I want us to standardize our terminology in the docs and how we communicate customer service. Let's treat this as a collaborative effort.

My take follows. I see this as a three-dimensional problem.

• Dimension 1 (collaboration): Is it individual, or shared?
• Dimension 2 (description): Is it a space, or an environment, or an allocation, or a quota, or an account?
• Dimension 3 (platform): Is it on GPFS, LTS, Cloud.rc, etc.?

I see "environments" as places where you do things. Allocations are for storage. I'm not a fan of "space" because it doesn't relate to the underlying technology, though it does have some inertia behind it.

I see "accounts" as a separate concept. An account is the digital representation of a person within the platform, and how we manage individuals authorization and authentication, and other person-specific features. Users have accounts, labs and cores do not have accounts.

I also see "Lab" and "User" and "Core" as the concept of "who", and also separate. Everything we do is ultimately about people and their research affairs. These terms imply dimension 1, collaboration (or lack of it), which is helpful. However, they also imply limitations on who we intend to help.

Putting all this together, these are the terms I think we should use.

• individual/shared Cloud environment for the underlying technology of OpenStack projects.
• individual/shared GPFS allocation for the underlying technology of GPFS quotas. That said, shared maps to /data/project/, and "project space" has a lot of cultural inertia behind it.
• individual/shared LTS allocation for the underlying technology of Ceph object storage "allocations" (proper name TBD).

Where is the inaccuracy?

All over the docs.

No documentation files or tests are named. Start by inventorying the repository docs for the listed terms and resolve the proposed vocabulary for Cloud/OpenStack, GPFS, and LTS/Ceph with stakeholders. Done means the terminology is agreed upon, applied consistently across the docs, and aligned with customer-service communication.