Large Datasets - How can I improve performance?

Hi GaryK the first question I would ask is when you say realtime what do you define as being realtime?

The reason for this is, is if you define realtime as being hourly or every 5 minutes, if you are looking for the optimal performance then importing your data into the Power BI Model will result in very quick response times when interacting with the data.

As v-haibl-msft has mentioned no actual data is stored in your Power BI Model. So if it is slow it is due to the Source system in your case your Oracle DB that is taking time to run the required queries and return the data back to Power BI. (And as you indicated it is doing 6 joins to underlying tables)

I would suggest possibly as a test, import all the data into a Power BI Model, and see if when using the Power BI Model with your slicers if it is quicker. If it is, then at least you will know where the bottleneck lies.

In my experience often people ask for realtime data, but unless someone is actually watching it as part of their job, it can be refreshed hourly, which is more than adequate.

Another option is if you only want to load specific data into your Power BI Model, you can make use of the Parameters. But note when you do this, every time you change the parameter you will have to re-import your dataset.