You have two very large databases supporting OLTP workloads which run on multiple small-blade style servers in a cluster and which require optimal latency for I...