Yes Eric, it is worth experimenting to find the performance issues before saying yes or no. I would triple (0r based on growth rate) the data in some large tables to ensure that this works when data grows in future.