DBclassify is built on an architecture that can scale to handle very large, multi-terabyte data volumes without impacting the performance of the database or BI applications.
The key elements of the architecture are:
DBclassify captures SQL statements in real time from Oracle's Shared Global Area (SGA). The non-intrusive SGA Sniffer enables DBclassify to capture all SQL statements running in the database, including those that do not originate from an end-user client—such local connections and stored procedures, packages, and functions. This “Listener” is designed to have a minimal fixed memory footprint and a statistically insignificant impact on both server resource consumption and database throughput.
DBclassify’s Packet Depot collects captured packets from the SGA Sniffer. It then sorts them by process and prepares them for the Packet Analyzer.
DBclassify’S Packet Analyzer processes the raw captured database SQL activity, extracting the logical session structure and the SQL requests. The processing is performed in a scalable multi-stage process. In the first stage, the Packet Analyzer uses database-specific parsing to process the raw messages. In the next stage, the Packet Analyzer assembles a model of all database interactions from the parsed database messages to create a representation of the database schema referred to as the Data Dictionary. This model incorporates both SQL statements and their session context.
DBclassify’s Repository maintains detailed usage history and database models. The Repository is instantiated as an Oracle database which can be directly queried by the user or by any database-oriented reporting tool (e.g., Business Objects, Cognos) in addition to DBclassify’s own reporting capabilities.
The GUI reports are generated through a Java application server.
DBclassify includes a rich GUI which can be accessed from any standard Web browser. The Management Console provides an intuitive set of controls for monitoring and configuring the DBclassify components.