Search and Match

The Scheduler and the MatchServers

The Agora SDK includes two server based software applications.  The Scheduler accepts requests to search and match an unknown identity from a workstation, and MatchServers working in parallel search though their own portion of templates, and return results to the Scheduler.  The Scheduler compiles the results and returns a response to the workstation.  The workstation has an Agora components called MatchClient that communicates with the Scheduler.  The Scheduler contains the entire database in an instance of SQL Server 2005.   The MatchServers hold all their allocated templates on computer memory.

Two dimensional scalability

The MatchServers may be grouped in clusters, called Logical Units.  When the database is small, one Logical Unit may have a small number of MatchServers and each contains its portion  of the templates on computer memory.   To maintain performance when the database grows, more MatchServers may be added to the Logical Unit.   Moreover, when the number of daily transactions against a growing database grows as well, more Logical Units with identical number of MatchServers may be added as well.  Each Logical Unit contains all the templates spread across the MatchServers.  This architecture allows the system to cope with a growing database AND with a growing number of transactions.