Batch insert operations within a single aggregate.

[schauder]

For storing non trivial aggregates many operations need to get performed.
Those should be performed using JDBC batch operations to minimise database roundtrips.

There will be issues for the different kind of operations:

• Inserts (this ticket)
• Batch delete operations within single aggregate #1160

Not all JDBC drivers return generated ids for batch operations.
Therefore we'll need a Dialect property describing if a dialect supports this.
If an id is to be generated by the database and the database does not support batch inserts with generated keys we need to fall back to sequential operations.

This builds on the verification of #1155
Note that this is about batching for a single aggregate, while #537 takes this concept across aggregates.