Absolutely. This was the original reason that we built this system. We found that we could get much much better performance when dealing with persistent data by putting the logic in the same process as the system which persists the data.