Understanding Scan Files

Data Platform

A scan file (referred to as an .xsf file) is a compressed format (i.e., gzip) of an xml file that an HP Universal Discovery agent generates on a server or workstation. The file is populated with all of that machine’s hardware and software inventory information. You can open a .xsf file with a tool such as winzip or winrar, and then open the .xml file contained inside for a readable list of inventory information on the corresponding server (one .xsf file equals one machine).

A comparison of the Pre-Data Platform Normalize and Post-Data Platform Normalize .xsf files was performed in order to document the changes Data Platform makes to the scan file during normalization and prior to loading to the UCMDB. The results of the comparison include:

•

In the <hwOSWMISoftwareFeatureDescription> and <hwOSServiceFileName> the file has quotes and apostrophes encoded (i.e., &quot; instead of “).

•

The file size is massively reduced after we run this process. i.e., the xml for Pre-Normalize files is 24mb and the xml for Post is 1mb. This is a huge savings in terms of storage and UCMDB processing due to a smaller file size for 10,000 scan files.

•

This happens because we strip out all of the <file> info from the Pre-Normalize files (all the xml tags for the dll, exe files, etc.). We also strip out the Partially Recognized apps (<partialapp> tags).

•

We put the normalized values into the application data section along with each normalized application within its own application tag. <applicationdata><application>.

•

We do not change the OS installed applications (<hwOSInstalledApps_value> tag in the xsf).