I would like to present the outcome of the technical use case research conducted this past summer/fall. (we just received public release approval on the paper) Back in July-Sept we conducted a series of interviews with members of the CPE Community in an attempt to better understand how CPE is being used.
A common theme taken from discussions at the Common Platform Enumeration (CPE) Developer Day (April 30th, 2008) was that differences in how and why CPE is being utilized across the community create conflicts among members of the CPE community, and thus create obstacles to CPE's success. In its non-competitive FFRDC role as moderator of CPE, MITRE has conducted a series of interviews with representatives of the CPE community (under a combination of implied and formal NDAs) and has performed an analysis of the different technical use cases for CPE. This white paper presents the results of that analysis.
* Four general use cases were identified including: Software Inventory, Network-Based Discovery, Forensic Analysis/System Architecture and IT Management.
* Universal agreement exists on the need to support the Software Inventory use case as a "must-have". No other use case was universally identified as a "must-have".
* Universal agreement exists that a) "concrete" names are needed to support the Software Inventory use case and b) these names can be created using current CPE 2.1 structures. But...
* Disagreement remains on how to best use CPE 2.1 to support the Software Inventory use case and the need for "abstract" names.