Skip to main content
Version: 3.2.0

Introduction to Automated Service Intelligence Machine Learning (ASIML)

The Interlink Software Automated Service Intelligence machine learning (ASIML) offering provides out-of-the-box support for machine learning techniques to automate alert correlation and alleviate the administrator from the need for manually coding alert management rules (e.g. policy-based temporal fuzzy matching, whitelist and blacklist processing).

ASIML Standard Alert Model (SAM)

Automatically correlate alerts into scenarios, dramatically reducing the volume of alerts the user must interact with. Inside each alert scenario exists one to many alerts that have been automatically correlated to that alert scenario. ASIML is controlled by the standard alert management policies defined in the ASIML dashboard. This is used to prepare and test the model for our alert and metric based machine learning.

SAM policies help to train the ASIMLa model algorithm on how to formulate its predictions. Each policy contains feature columns and match weightings.

Three default SAM policies are provided out-of-the-box, Source, Description and IPCheck. the following settings for the feature columns and match weightings are used:

PolicyFeature ColumnAlgorithmWeighting
SourcemcnameSimilarity100
Source_domainSimilarity75
DescriptionmcnameSimilarity100
DescriptiontextSimilarity75
IPCheck_asiml_ipaddrSubnet24

ASIML Field descriptions

FieldDescriptionExample
asiml_automateOn the parent alert, the value "Ticket". To enable you to use this as a trigger when creating incidents.Ticket
asiml_countOn the parent alert, the number of child alerts under the specific scenario.6
asiml_ipaddrThe IP address of the alert if it’s in the alert.192.168.1.3
asiml_labelThe unique Scenario Name on the parent alert. The policy name on the clustered alerts.SN00000008
asiml_precisionPercentage of how precise it matches to the ASIML policy.89.4736856
asiml_scenarioThe BES AlertId of the parent alert0006736478
NotesFull details of all policies which have clustered on the parent alert2019-03-08 11:35:45 ASIML: Associated Alert 0006736478:Description:89.4736856% - mcname=ISS text=app2 M Node down

Where Next