PLATFORM

PAF’s Platform Architecture

PatternBuilders’ Analytic Framework is made up of the three layers:

• Analytics Server—the logical components that form the basis of the Framework.
• Direct Integration—where PatternBuilders’ or third party modules can integrate directly into the analytics application.
• SOA Interface—provides access to all “public” interfaces of the analytics server and modules in the direct integration layer for a particular implementation and is the most common level where customization and integration occurs.

The following diagram shows PatternBuilders’ Analytic Framework architecture. 

The framework is designed to be “storage agnostic.” In other words, the analytics it produces, as well as the underlying data, can be stored in a traditional relational database such as SQL Server or Oracle, some Object Databases, a journaling file system, or in a Netezza advanced Data Appliance.

About the Analytics Server Layer

The analytics server is made up of six logical components that form the foundation of the Framework. These components, or modules, can be located on separate servers (fan out) depending upon what the network requires. The components are:

• Data Loading Server. This server loads data from a relational database, CSV and XML files for analytic calculations.
• Analytics Calculation Engine. Using an advanced statistics library, the Analytics Engine calculates analytics (KPIs) based on incoming data.
• Metadata Server. The metadata server stores data about the application domain and allows PAF user interfaces to be easily configured without writing code.
• Geo-Coding/Mapping Server. This server stores the geo-codes for any physical location, enabling geographic queries and mapping mash ups.
• Business Rules Server. Enables business process automation across all KPIs in the system.
• Graphical Query Server. Enables easy hierarchical queries and delivers the fast return of results for any KPI in the system

About the Direct Integration Layer

The direct integration layer enables modules to integrate directly with the Analytics Server, whether the modules are developed by PatternBuilders or third parties. The pre-built PatternBuilders’ modules that make up this layer are:

• Email and Microsoft Office Application Integration Modules. These modules directly integrate Microsoft Office and email applications, enabling the use of both from within the analytics application.
• Auto-Generated User Interface. The Framework automatically generates a default user interface that includes both an administration module and a metadata configuration editor. The editor can be used to configure default user interfaces without coding.
• Direct API Integration. This API enables complex integrations with the Framework internals. The majority of applications do not need to integrate at this level.