Usecases

Using Data analytics and Machine learning to model Traffic Prediction, Traffic Classification, Anomaly Detection and Fault Prediction, we support a range of usecases that are common in network and security operations.

Some of the more common ones are listed below. We can support many others and we’d love to hear about your specific usecase(s).

Traffic Prediction

Network traffic trends are used to predict bandwidth utilization and latency. This enables improved capacity planning, and QoS provisioning

Read More

Anomaly Detection

Network traffic anomalies cause network operations to deviate from normal behavior. This enables detection of impending failures of security attacks

Read More

Fault Prediction

Network faults can be predicted and classified based on prior knowledge. This enables proactive mitigation of failures and reduces time to repair

Read More

Traffic Prediction

Data Required (any one)

  • Netflow data
  • Packet Header data
  • SNMP Interface statistics

What is predicted ?

Short-term and long-term changes in the amount of network traffic.

How to use the predictions?

Based on the prediction of bandwidth saturation, the network engineer can proactively suggest a bandwidth upgrade or provision QoS.

Predictions can also be used to determine more accurate thresholds for bandwidth and usage metrics.

Anomaly Detection

Data Required (any one)

  • Netflow data
  • Packet Header data

What is predicted ?

Traffic deviations from the baseline.

Traffic behavior changes (e.g., one source IP address contacting multiple destination IPs within a short time frame).

How to use the predictions?

With predictions based on detected anomalies, network engineers can further investigate potential faults or threats that can cause a problem in the future.

Fault Prediction

Data Required (any one)

  • Netflow data
  • Packet Header data
  • SNMP TCP/UDP statistics

What is predicted ?

While anomaly detection predicts anything that deviates from baseline behavior, fault prediction provides link, device and connectivity faults and performance degradation.

How to use the predictions?

Using the predicted faults and their probability, network engineers can plan ahead for maintenance and proactively fix any faults or performance degradation flying under the radar.

Predictions can also be used to determine more accurate thresholds for alarms and performance metrics.