Improve computer models management between Fusion Inventory, GLPI and automatic provisionning tools

I have built out of GLPI, FusionInventory, CustomFields, Microsoft MDT (Deployment tool) and a home made database synchronisation tool something that is close to SCCM at very low cost.

I plan to relase it as soon as it is clean enough to be shared.

Basically, I use the data from GLPI to feed the MDT database. This enable automatic OS deployment just pressing "F12" at boot (PXE boot to MDT). Our technicians just choose the OS version, the machine name, the additional applications (MS office, OpenOffice....) in GLPI and that's it.

In order to manager hardware specificities, I rely on the computer model to set some default, model based values. This used to work fine until the lastest update (21) in FusionInventory. A lot of machines where identified with a technical code for the model rather than a commercial name. This code is what MDT relies on to identify a model (plus, several models with slightly different hardware may often the same commercial name). The output on screen is much nicer (20C600JJFR is now ThinkPad Edge E540) but MDT is broken.

Therefore my suggestion is:

- Have GLPI manage both the commercial name and the technical code for computer models.

- Add one column to the glpi_computermodels table for code. Make this column indexed unique while names can be non unique.

- Show both code and name on computer description page.

- Adapt Fusion Inventory import code to store both informations.

Indeed, there will be an issue when two different versions of FI will report a different name for the same code. I guess the rule should be:

- If code and name reported are the same and conflict with an existing model where they differ, keep the one that differs as this is probably more up to date.

- If code and name reported are different, then update table. Last to talk wins!