Nuances of the File Classification Infrastructure

Dealing with rule overlap and rule re-application

One of the interesting things that can happen with file classifications in Server 2008 R2 is that you may encounter some “rule overlap.” For example, if two rules apply to the same file, and one sets a given property to the third value in an ordered list, and the other rule sets the same property to the second value, the second value will be the winner. In similar fashion, if one rule sets a property to a Boolean “yes” and another rule sets the same property to “no,” the result will be “yes.” However, you can override this “aggregation” behavior on the Evaluation Type tab of the Additional Rule Parameters dialog box in the File Server Resource Manager (FSRM) console; for example, you can specify “Overwrite the existing value” instead of “Aggregate the values”.

In any case, it’s a good idea to imagine the possible scenarios when you’re designing your rules in the new File Server Resource Manager (FSRM) console, so they behave as you intend.

Another thing to consider is that by default, the FCI system will not apply any new rules to files that have already been classified with a given property. If you want to change this behavior and force Server 2008 to re-examine previously classified files according to any new or changed rules, you should check the “Re-evaluate existing property values” box on the Evaluation Type tab mentioned above.

One of the things I’m going to be interested to see on production systems is how quickly files actually get classified. Sometimes these technologies sound pretty interesting but they can be fairly slow to be executed on real systems (SMS, anyone?). If any of you already have some experience with FCI and would like to comment on its performance on your network, please feel free to share your thoughts.

• Microsoft
• fci
• file classification infrastructure
• file server resource manager
• fsrm
• server 2008 r2