Snow Atlas introduces a more advanced and layered approach to device identification compared to its on-premises counterpart. At the core of this improvement is the Device Identity Service, which intelligently processes incoming data and applies a hierarchy of matching rules to determine whether to associate the data with an existing device record or create a new identity.

How it works

The Device Identity Service follows a structured process to ensure accurate and consistent device identification:

1. Extract and Normalize Predictor Fields:

• Hostname
• Fully Qualified Domain Name (FQDN)
• BIOS Serial Number
• Established IDs (like cloud provider IDs)

2. Apply Matching Rules in Priority Order:

• Match by Established ID
• Match by FQDN
• Match by BIOS Serial Number + Hostname
• Match by BIOS Serial Number + Hints (like MAC address, IP address)
• Match by BIOS Serial Number only (if FQDN is unavailable)
• Match by Hostname only (if FQDN is unavailable)

3. Determine Final Identity:

• If multiple matches are found, the highest-ranked match is selected
• The matched identity is returned to the inventory system

Matching criteria

The following data types are used in the matching process:

• Established IDs:

• Amazon EC2 Instance ID
• Azure VM ID
• Google Cloud Computer ID
• Linux Asset Tag

• Primary Fields:

• Hostname
• Domain
• BIOS Serial Number

• Hints (Tie-Breakers):

• • MAC Address
  • IP Address
  • Source Name
  • BIOS Manufacturer
  • Processor Manufacturer
  • Processor Speed

This intelligent matching logic ensures that each device is uniquely identified, prevents duplication, and maintains consistent tracking across systems.