File Classification Pipeline

The File Classification Infrastructure (FCI) uses a five-step pipeline to classify and manage files. The first four steps classify the files, and the fifth step uses policy to manage the files.

Step 1, Discover Files, assembles a list of files for classification based on a classification rule search scope, such as a particular volume or folder. It then creates an in-memory property bag for each file. The property bag is used to hold sets of “Name=value” classification properties for a given file.

Step 2, Extract Classification, uses components called storage modules to retrieve previously stored classification data from documents created in applications such as Word or stored in libraries in SharePoint. FCI includes storage modules for Office 2003 and 2007 document files, and a generic system storage module extracts the file properties stored in a file’s alternate data stream. (In the NT File System, all data is stored in a file’s main unnamed data stream, but developers can also create alternative data streams to contain additional information, such as file classification properties.)