Well, you have a lot of moving parts in your system.
Spark only supports SQL Server. Furthermore, to benefit from the code-generating facility your data model must adhere to some conventions (primarily about the primary key: they must be int identity fields that are named Id).
On the other hand Spark does support the use of multiple databases, but, again, only with SQL Server and each database must adhere to the conventions.
A lot can be done to adjust Spark to your particular environment. However, I am not sure that would be the right decision. With your list of legacy and other tools you may be better of with Entity Framework, Nhibernate,or some other ORM.
Hope this helps.
Jack Poorte, Jul 08, 2014