300-445 Question 61
Select 3An enterprise network operations team is implementing a network assurance solution to monitor performance across key links in the network. As part of this process, they need to set a metric baseline to identify deviations and ensure optimal performance. Which of the following are key considerations when setting a metric baseline?
- A
Collect performance data during a period of normal network operations.
- B
Use data from peak traffic periods exclusively to set the baseline.
- C
Account for seasonal or time-based variations in network traffic.
- D
Set the baseline metrics once and avoid revisiting them to ensure consistency.
- E
Use historical data to compare against live data for anomaly detection.
Show answer and explanation
Correct answers: A, C, E
Explanation
To set an effective metric baseline, data must be collected during normal operations to reflect typical behavior, and seasonal or time-based variations must be accounted for to ensure the baseline is accurate and adaptable. Additionally, using historical data for comparison enhances anomaly detection. Neglecting these aspects or basing the baseline solely on peak traffic periods would result in an inaccurate and ineffective baseline.
- A. Correct.
Collecting performance data during normal network operations ensures the baseline reflects typical network behavior, making it easier to identify deviations accurately.
- B. Incorrect.
Using data exclusively from peak traffic periods would not provide a balanced baseline, as it fails to account for normal or low-traffic periods.
- C. Correct.
Seasonal and time-based variations in traffic patterns can impact network performance, so they should be considered to create an accurate and adaptable baseline.
- D. Incorrect.
Setting baseline metrics once and ignoring them over time is a poor practice, as networks evolve and traffic patterns change. Baselines should be revisited periodically.
- E. Correct.
Historical data provides a valuable reference for understanding typical performance, enabling effective anomaly detection when compared with live data.