7.1 AI/ML Analytics Overview

7.1.1 DNAC AI-Driven Network Capabilities

+------------------------------------------------------------------+
|                    DNAC AI/ML ARCHITECTURE                        |
+------------------------------------------------------------------+

                    +---------------------------+
                    |      AI/ML Engine         |
                    |   (Cloud + On-Premises)   |
                    +---------------------------+
                              |
        +---------------------+---------------------+
        |                     |                     |
+---------------+    +----------------+    +----------------+
| Anomaly       |    | Predictive     |    | Automated      |
| Detection     |    | Analytics      |    | Remediation    |
+---------------+    +----------------+    +----------------+
        |                     |                     |
+---------------+    +----------------+    +----------------+
| Baseline      |    | Trend          |    | Self-Healing   |
| Learning      |    | Forecasting    |    | Actions        |
+---------------+    +----------------+    +----------------+
        |                     |                     |
        +---------------------+---------------------+
                              |
                    +---------------------------+
                    |     Network Telemetry     |
                    | (SNMP, Syslog, NetFlow,   |
                    |  Streaming Telemetry)     |
                    +---------------------------+

7.1.2 AI/ML Use Cases

Use Case	Description	Benefit
Anomaly Detection	Identifies abnormal patterns in network behavior	Early issue detection
Root Cause Analysis	Correlates events to identify underlying issues	Faster MTTR
Predictive Insights	Forecasts potential failures before they occur	Proactive maintenance
Trend Analysis	Identifies long-term patterns in network usage	Capacity planning
Client Experience	Analyzes user connectivity patterns	Improved satisfaction
Security Analytics	Detects suspicious traffic patterns	Threat identification

7.1.3 Machine Learning Models

ML_Models_in_DNAC:

  Baseline_Learning:
    Type: Unsupervised clustering
    Purpose: Establish normal behavior patterns
    Training_Period: 14 days minimum
    Updates: Continuous (rolling window)

  Anomaly_Detection:
    Type: Statistical + ML hybrid
    Purpose: Identify deviations from baseline
    Sensitivity: Configurable (Low/Medium/High)
    False_Positive_Rate: <5% target

  Predictive_Failure:
    Type: Time-series forecasting
    Purpose: Predict device/link failures
    Horizon: 24-72 hours ahead
    Confidence: >85% accuracy

  Root_Cause_Correlation:
    Type: Graph-based reasoning
    Purpose: Correlate multi-source events
    Data_Sources:
      - Syslog events
      - SNMP traps
      - NetFlow anomalies
      - Client connectivity

7.1.4 Enabling AI Analytics

# DNAC Configuration for AI Analytics

# System > Settings > AI Analytics

AI_Analytics_Settings:
  Cloud_Connectivity:
    Enabled: Yes
    Endpoint: analytics.cisco.com
    Data_Sharing: Telemetry only (anonymized)

  Baseline_Learning:
    Enabled: Yes
    Learning_Period: 14 days
    Sensitivity: Medium

  Anomaly_Detection:
    Enabled: Yes
    Categories:
      - Device health anomalies
      - Client connectivity anomalies
      - Application performance anomalies
      - Security anomalies
    Alert_Threshold: Medium confidence

  Predictive_Insights:
    Enabled: Yes
    Forecast_Window: 48 hours
    Notification: Email + Dashboard

---