Telecom

OptOSS AI Product Owner’s Cheat Sheet

OPT/NET
·
September 11, 2024

OptOSS AI is a sophisticated tool which enhances the management and optimisation of telecommunication networks. Understanding how to navigate and leverage its features is crucial to ensuring that your team can quickly identify, analyse, and resolve issues. Together with our Product Owner Vasyl, we’ve collected  some tips and tricks to help you get the most out of OptOSS AI:

1. CT Scan View: Your First Line of Defense

The CT Scan View is one of starting points for identifying anomalies within your network. This feature allows you to visualise all your anomaly data across your devices over a specified time period in 2D & 3D, making it easier to detect unusual patterns and potential issues.

  • Select a 12-Hour Period: Begin by setting the time range to 12 hours. This interval provides a balanced view, capturing enough data to reveal trends without overwhelming you with information.
  • Filter for Anomalies Only: To focus on critical data, apply a filter that displays only anomalies. This step helps in quickly identifying deviations from normal operations that may require attention.
  • Use the Highlight Field: The highlight feature enables you to pinpoint anomalies related to specific devices or locations. By narrowing down the focus, you can target problem areas more effectively.
  • Navigate Through Time: Utilise the navigation buttons or hotkeys to move backward or forward in time. This functionality is particularly useful for tracing the evolution of an issue and understanding its impact over time.

2. Clusters View: Prioritising and Classifying Issues

The Clusters View is a powerful tool for prioritising and classifying network issues based on the frequency and severity of anomaly clusters.

  • Select a 24-Hour Period and Sort by Amount Anomalies: Begin by setting the time range to 24 hours and sorting clusters by the number of anomalies. This approach provides a prioritised list of issues within the network, helping you focus on the most frequently occuring repeating anomalous sequences (problems) first.
  • Filter by Priority: Apply filters to understand how often a particular issue has occurred over the selected period (filter by region, severity, frequency, tags, etc). This filtering helps in identifying recurring problems that may require immediate attention.
  • Filter Unclassified Clusters: Shift your focus between repeating and new clusters which have not yet been classified. It's crucial to minimise the number of unclassified clusters, as this reflects your level of awareness and control over issues within the monitored networks. (TIP: If you’ve previously classified similar clusters, check them to quickly identify and automate the classification of these new anomalies.)
  • Check Script Assignment: Ensure that the appropriate script is assigned to target clusters. Proper script assignment is vital for automation, ensuring that the correct actions are taken automatically when new anomalies of the respective cluster are detected.

3. History View: Investigating Past Incidents

The History View is essential for deep-diving into previous incidents for forensic analysis, allowing you to understand what went wrong and how to prevent similar occurrences in the future.

  • Select Devices from the Last Incident: Start by identifying the devices involved in the most recent incident. This selection will provide a clear picture of the affected areas.
  • Adjust the Time Range: Set the time range to include moments before the incident. This adjustment helps in identifying precursor events or conditions that might have contributed to the issue.
  • Navigate to Events: Once you’ve identified a suspect anomaly that likely caused the incident, you can drill down into the specific events (for example the raw Syslog data).

4. Dashboard: Monitoring at a Glance

Create your own dashboard with your favourite widgets. Here’s some of our Product Owner’s favourites:

  • Anomaly Trends Widget: Use cluster tags to track the evolution of issues over time. This widget helps in identifying patterns, including the initial cause (patient zero) of recurring problems.
  • SUN Widget: The SUN Widget highlights areas with the most significant issues. Cross-reference these insights with detailed data from the CT Scan View to gain a deeper understanding of the problems.

5. AI-Analytics: Advanced Cluster Management for Advanced Users

AI-Analytics is where you refine and optimise your cluster management strategy, ensuring that your network operates at peak efficiency.

  • Cluster Performance, Dig-In View, Delete cluster: Identify clusters with low homogeneity and numerous anomalies. By splitting these clusters, you can create smaller, more homogeneous groups that are easier to manage and optimise. After making adjustments, check the results and classify the new clusters accordingly.
  • Similar Clusters View: Sometimes new clusters form which require the same treatment as an existing classified cluster. You can apply existing classifications to unclassified clusters when they have a high similarity score to existing clusters. The similar clusters view allows you quickly, appropriately classify new clusters using a similarity matrix.
  • Enable Auto-Classification: Once you've classified clusters, enable auto-classification to automatically apply the same classification to new clusters above a certain similarity score threshold. Just make sure it’s a high score like 95% to avoid bundling different anomalies!.
  • Restore Classification View: Check a list of saved auto-classifications and manage them if needed. You can turn these classifications ‘ON’ or ‘OFF’, giving you control over the classification process and helping you to automate new cluster classification.

Navigating OptOSS AI effectively requires a systematic approach: start with a helicopter view of your network, zoom in on specific cases, investigate anomalies, automate solutions, and repeat the process. By mastering these steps, you’ll enhance your ability to manage and optimise your network.

Ready to see OptOSS AI in action? Watch the demo video:

Contact us to learn more about how we can help you transform your network management or for any other inquiries!