There are some optimizations we can make around validation of replicated data.  In particular, we could skip syntax checking for replicated operations since the syntax checking of the new values should have been performed on the master that initially received the operation from a client.  This optimization should help improve performance of replicated operations.

We should add the ability to configure if syntax checking should be skipped for replicated operations to allow flexibility for different deployments.  We also may want to think if there are other things we could skip around replicated operations, such as schema checking.