When multiple data sources being set in a system, there should not be an assumption that a subscriber that is not found in one data source will be available in the next one. Each data source is expected to have a different set of subscribers (or keys) and filters on the
MPE that are used to direct queries to a specific data source. The only scenario is were the provisioned subscriber base is expected to be replicated among the servers within a data source. If you have multiple data sources, the following are the functionality scenarios supported.
- Primary and Secondary role data sources
Primary and Secondary roles are used to designate a data source to target with the first query. A key retrieved from the first server (Primary) is used to query the Secondary server. If the Primary server query times out, the Secondary server is not queried because of the dependency.
- Primary Data sources configured for Merge Results functionality
The Primary Data sources are queried simultaneously to retrieve different sets of data about a subscriber which is then merged for a complete profile. There is not a question of timeout and the try next data source as dependency does not exist as in scenario 1.
- Multiple Primary data sources with different provisioned subscriber base
Filters are used to direct queries to specific Data Sources. There is not an automatic query of the next data source after timeout because filters for each data source are different. the
MPE can still query across the data source boundary using the priority order in the GUI data source table.
Important: There must be an explicit UNKNOWN SUBSCRIBER response from the first data queried for this to occur. A timeout or no response from first data query will not trigger the lookup of next data source