Content Classification Service Integration Guide › Introduction › Deployment Architecture

Deployment Architecture

When integration with the CCS is enabled, external applications call the CCS web service when they need a document to be classified. The CCS analyzes and classifies the document, or looks up a cached classification for the document, and returns the result to the external application. The following diagram shows an example integration with CA SiteMinder.

A user (1) attempts to access a document in a SharePoint library (2).

SiteMinder intervenes (3) and passes the document URL (4) to the CCS. In this example, multiple CCS servers are deployed to maximize performance. These include a primary CCS server (5a) and two secondary CCS servers (5b).

The CCS examines the document to determine whether it has been previously classified.

• If the document has been previously classified, the CCS retrieves that classification from the shared Classification Cache (6), hosted in CA Directory.
• If the document has not been previously classified, a CCS policy engine (7) analyzes the document's text content to determine whether the document matches a predefined classification. The actual policy, including the necessary triggers and document classifications, is stored on the CMS (8) on the primary CCS server. If policy analysis confirms that the document matches a predefined classification, the results are stored in the Classification Cache (6),

If the CCS can confirm that the document matches a predefined classification, it returns that classification to SiteMinder (9). In the example above, the document matches a Personally Identifiable Information (PII) classification. SiteMinder can then deny or grant access to the document, depending on how its SharePoint access rules are defined. You can customize the default classifications in the iConsole (10).

You can use the CCS preclassification agent (11) to scan and classify documents offline, so avoiding the need to analyze document content in real time. The resulting classifications are stored in the Classification Cache (6). You manage preclassification scans in the Administration console (12).