I can answer the first question:

You can define data storage by setting the `path` parameter for tables. The "storage path" in pipeline settings will then only hold checkpoints (and some other pipeline stuff) and data will be stored in the correct account (if you leave pipeline "storage path" blank, default `dbfs:/pipelines/<pipeline-id>` will be used).

Every delta live table is created in metastore - so schema/table grants should be used to manage permissions per layer. We use schemas to separate layers.

My wish is that the schema would also be part of the table property and overwritten, not only the pipeline setting.