DS deployment architecture, which globals should be mapped where?

To deploy a DeepSee solution, the docs recommend that you define a namespace on the reporting (mirror) server, and "define mappings to access the application data, application code, DeepSee cube definitions, and DeepSee data on this server".  (http://docs.intersystems.com/ens20152/csp/docbook/DocBook.UI.Page.cls?KEY=D2IMP_ch_overview#D2IMP_overview_architecture)

This implies that for an ideal deployment architecture, globals should be split into four separate databases (app data, app code, DS cubes, DS data).  How exactly should the DeepSee-related globals be split?  A table in the docs outlines all of the key globals (http://docs.intersystems.com/ens20152/csp/docbook/DocBook.UI.Page.cls?KEY=D2IMP_ch_setup#D2IMP_setup_globals_alt), but is there an optimal way to split those into 'code' and 'data'?

Thanks,

Steve