Embedded memory consumption - Newbie question

[tommy_g]

I am contributing to a product which will run as an embedded application. I understand memory consumption when developing for a Pro or PPU environment (mult-tenant). I know in general about capacity settings, but have no control over them. 

If I have a moderate DM which is used by several dozen users, do they all share the same in-memory, column store database? (assume 95% import mode)

If I have a very large DM (inherited from an earlier failed developer) what are the relevant best practices to re-factor. My first instinct would be to modularize via Dataflows, but I do not know how an embedded plan responds. I've been told that the parent portal application ONLY allows one dataset, hence the very large DM. I think this is nuts, contrary to all my experience. However, I need to cover what I don't know, and maybe even what I don't know that I don't know about embedded.

Something of an open-ended question for now, I need expert guidance on how to frame more specific questions.

[GilbertQ · Accepted solution]

Hi @tommy_g 

You can have multiple datasets when using Embedded. They will all be consumed within the embedded capacity.

And yes you are right that the moderate DM can be used by a lot of people because the single DM is shared amoungst all the users.

If you have more details and very specific questions I would suggest looking for a consultant to assist you. While the commuity is here to assist you with questions and answers I would not suggest asking questions for which a consultant or company should be helping you with.

[tommy_g]

The only downside of your recommendation to look for a consultant is that I AM the consultant.

Talking to myself is rarely sufficient 😃