Automatic inventory without dictionaries is a polluted CMDB. Each agent reports data with different spellings, generating duplicates and inconsistencies. Dictionaries solve this automatically.
The problem
When GLPI Agent collects data from hundreds of machines, it is common to find:
- Manufacturers: "Dell Inc.", "Dell", "DELL", "Dell Technologies"
- Models: "Latitude 5520", "LATITUDE 5520", "Dell Latitude 5520"
- Software: hundreds of Windows KBs mixed in with real software
- OS: "Windows 10 Pro", "Microsoft Windows 10 Professional", "Win10Pro"
Without standardization, reports and dashboards become useless.
Available dictionary types
Go to Administration > Dictionaries:
- Software: rename, group, or reject imported software
- Manufacturer: unify manufacturer names
- Models: computer, monitor, printer, phone, network
- Types: computer, monitor, printer, device, network, phone
- Operating system: OS, version, service pack, architecture
Practical examples
Unify manufacturers
Criterion: Manufacturer contains "Dell" → Action: Assign manufacturer = "Dell"
Result: "Dell Inc.", "Dell", "DELL" → all become "Dell".
Filter Windows KBs
Criterion: Software starts with "KB" → Action: Reject (do not import)
Criterion: Software contains "Update for Windows" → Action: Reject
Standardize OS
Criterion: OS contains "Windows 10" AND OS contains "Pro" → Action: Assign OS = "Windows 10 Pro"
Best practices
- Always test rules in a staging environment before production
- Start with manufacturers and software – they are the most polluted
- Run the dictionary replay after creating new rules to apply them to existing data
- Review dictionaries after each large inventory import
Next step
With standardized data, your CMDB will be ready to generate reliable reports and dashboards. Also check the asset lifecycle for complete asset management.
