09/22/2022, 8:36 PM
I'm still really early in understanding SpiceDB. I was wondering how people think about their application design, and "datastores" in general, that "live along side SpceDB". For example, we use DynamoDB as our primary datastore. I've read up on the ZedTokens. I think i get that part of it..... From an operational POV, it feels like the SpiceDB datastore, needs all the same "-isms" as my Application Datastore. Operational expertise, Backups, HA, multi-region, etc. Maybe some specific questions: - Do people write their "list of editors" for an object, also in their Application datastore? Or is it only stored in SpiceDB? eg, do you store enough data in your Application that you could "reconstruct" the relationships? - At some scale the requirements for spiceDB are probably different scaling characteristics than my primary application -- are people deploying them as parallel systems, or are people re-using their existing database and also running SpiceDB on it?