Praetor is almost linearly scalable by adding more CPUs within the recommended cluster.  Its architecture is designed to use one central SQL Server to log all message summary information and to hold the common Bayesian token database.

As message volume grows, capacity can be increased by adding CPUs to perform filtering.