File Classification Infrastructure and Server 2008 R2

How to go beyond built-in file properties

Remember WinFS – Windows Future Storage? Once upon a time, this was a planned radical enhancement to the Windows file system that was going to be part of Vista. It turned out to be far too ambitious, far too Microsoft-centric, and far too slow to see the light of day; but the basic idea was sound. NTFS is a good file system, but it just doesn’t let administrators and users describe files in adequate detail and with sufficient flexibility. As the number and type of files on our servers continues to increase, we need better tools to manage those files.

File management consists of a wide variety of tasks. You have to back up your files, and from time to time, restore them. You have to secure your files. You have to archive them, in one or more stages, based on expiration criteria. You’d like to do some reporting so you have a handle on what types and sizes of files are where. You need to perform capacity planning. Of course we’ve been doing these things all along, but it sure would be easier if we could extend the file properties architecture to be customizable to our needs. The mere physical location of a file in a folder is not enough for us to make many of the file management decisions we need to make. In fact, depending on your network, the folder location may be next to irrelevant!

In Windows Server 2008 R2, Microsoft has extended the File Server Resource Manager console with a new node titled “Classification Management.” This is part of what the company calls FCI, File Classification Infrastructure. (It’s an “infrastructure” because you can interact with it via the FSRM console or via PowerShell.) The idea behind FCI is to let organizations define file properties that will be useful to them, build rules to apply these new custom properties, and perform file management operations that take the new file properties into account.

The first step is to figure out what new classification properties you might like to create. (Remember that these are operating-system level properties, so they have the virtue of not depending on any particular file type or application.) You’ll create these properties in the FSRM console using the “Classification Properties” node under “Classification Management.” The allowable properties have a decent variety of data types, including string, list (multi-string), ordered list, Boolean, date/time, numeric, and multiple choice. Once you’ve created the properties you need, such as “confidential” or “never expire” or “trade secret” or “relating to the Widget Project” or whatever, then you can create rules to apply those classification properties to existing files – a subject we’ll look at tomorrow!

• Microsoft
• fci
• file classification
• file classification infrastructure
• file management
• server 2008 r2