OpenX has support for the Distributed Statistics. Distributed Statistics Architecture can be used for the really large setups where single core database is not enough. In Distributed Statistics architecture there will be a number of web servers. Each web servers runs a slave database (partial replication of a core database) and logs its own statistics to the local database, and perform a partial statistics summary locally. These all data from the all slave databases are pushed to the core database, where the remaining statistics generation that requires a complete view of the raw data can be performed.