Create explanatory model for new storage platforms

What would you like to see added?

The new storage platforms will have three parts. My current mental model is memory and storage in a computer, but we need to translate this for non-technical audiences.

• What we call "GPFS" or "Cheaha Storage"
  • Has a component with the technical name GPFS. We can think of this as a cache that stores the most frequently used data.
  • Has a component with the technical name Ceph Pool. We can think of this as the system RAM.
  • Data is moved into the cache when it is used, and moved back to RAM when it has not been used recently, and something else takes its place.
  • Data coming from "RAM" to "cache" may take some time, causing a small amount of latency the first time the data is accessed. In most cases this shouldn't be an issue.
  • Transfers happen automatically.
  • We won't be doing manual transfers.
  • Data must be in the cache to be processed on Cheaha.
• What we call "Long-Term Storage" or "LTS".
  • We can think of this as a hard drive. Data has to be moved from the hard drive into cache to be processed.
  • Transfers are not automatic and must be initiated by researchers intentionally.
  • Transfers can be partially automated as part of a processing pipeline or workflow.