Thank you Sergey.

Just a follow up on #2. If we maintain our surrogate keys in integer format and manage the foreign keys ( use them for joins in queries etc... ) we can just let Sql Server create its own guid columns ( as I understand one per table to identify that record ) and it will manage them without our input?

We rely heavily on meta data definitions to allow us to build queries at runtime and we could just ignore the guid columns when pulling field lists right?

I am leaning heavily to advising the team to stay with the integer keys and let Sql handle its own guids just for simplicity and also to keep from having to convert the integers to guids and establish the foreign keys links.

Thanks again.