This article provides an overview of the device matching logic and instructions on how to configure matching levels and manage records that require manual review.

Due to the broad scope of infrastructure discovery, duplicate items may occur. To consolidate records and maintain a single source of truth, a proprietary matching algorithm is applied to all incoming discovery data.

When a new discovery is performed, the algorithm compares various device attributes against existing records. If the calculated match score meets a specific threshold, the existing record is updated; otherwise, the device is flagged for review or created as a new asset.

Prerequisites

Review the following requirements to ensure your discovery jobs are optimizing for record accuracy.

You need the following before the installation:

• Administrator privileges to access Global Settings.

• Successful completion of at least one full discovery job (Windows, *nix, or Cloud).

• Understanding of your environment's naming conventions and hardware consistency.

Configure the device matching level

Choose a matching model that aligns with the quality of data in your network.

1. Navigate to Tools > Global Settings.

2. Locate the Miscellaneous section.

3. Select one of the three available matching models:

   • Classic: Performs a direct check of the Serial Number, UUID, and Name (in that order). Any match triggers an update.

   • Moderate: Uses a weighted scoring system. A device is considered a match if it achieves a score of 20 points or higher.

   • Conservative: Uses the same weighted system but requires a higher threshold of 30 points for a match, reducing the risk of false positives.

Enable enhanced device matching

Improve accuracy in complex environments where different devices may share the same hostname.

1. Ensure the Device Matching Level is set to either Moderate or Conservative.

2. Toggle the Enhanced Device Matching option to On.

3. This setting uses additional heuristics to differentiate between devices with duplicate names and validates potential matches discovered through Application Dependency Mapping (ADM).

Device attributes used in matching

The algorithm weights properties based on their significance to ensure reliable identification. The attributes are categorized into four levels:

Managing unmatched devices

A device must have at least one attribute match to initiate a score calculation. If no common attributes are found, the system treats the device as a new record.

Handle unprocessed records

While the advanced matching model prevents false positives, it can result in "unmatched" flags if the score falls below the required threshold (20 or 30 points).

1. Navigate to Discovery > Unprocessed Device Records.

2. Review the list of flagged devices that could not be matched with high certainty.

3. Manually merge these records with existing devices or approve them as new entries to clear them from the unprocessed table.