Advanced machine learning techniques for internet outage detection
The present systems and methods provide a user performance monitoring solution that enables the monitoring of application and device performance from the end user's point of view. The present systems and methods help Information Technology (IT) personnel to ensure the quality of digital experience across the enterprise. The present system is adapted to collect telemetry data from devices relative to the performance of all tiers of Internet Service Providers (ISPs), create a baseline of the performance of the ISPs based on a plurality of metrics and the collected telemetry data, train a Machine Learning (ML) model to assess blackout and brownout prediction accuracy at different performance values for the metrics, and identify a blackout or brownout, wherein a blackout or brownout is identified when real time performance is worse than the performance values identified by the model.
1. A non-transitory computer-readable medium comprising instructions that, when executed, cause a processor to
receive, at a cloud-based system, telemetry data from a plurality of connector applications associated with a plurality of user devices, wherein the telemetry data is collected at the plurality of connector applications, the telemetry data being relative to performance of all tiers of Internet Service Providers (ISPs) from a point of view of the plurality of user devices;
generate baselines of the performance of the ISPs based on a plurality of metrics and the collected telemetry data, wherein the baselines are generated over a plurality of geo granularities;
train a Machine Learning (ML) model to assess blackout and brownout prediction accuracy at different performance values for the metrics; and
identify a blackout or brownout and determine a severity of the blackout or brownout, wherein the blackout or brownout is identified when real time performance is worse than the performance baselines identified by the model, and wherein the severity of the blackout or brownout is based on a number of users affected and a geographic size of an affected area.
2. The non-transitory computer-readable medium of claim 1 , wherein blackouts and brownouts are detected by computing distance of current metrics from baselines.
3. The non-transitory computer-readable medium of claim 1 , wherein the collecting telemetry data comprises performing traceroutes from the plurality of user devices to frequently visited destinations.
4. The non-transitory computer-readable medium of claim 3 , wherein the traceroutes are performed between data centers set up across geographic locations.
5. The non-transitory computer-readable medium of claim 1 , wherein the plurality of metrics comprise Page Fetch Time (PFT), Time To First Byte (TTFB), probe error rates, latency, packet drops, and route stability.
6. The non-transitory computer-readable medium of claim 1 , wherein a blackout or brownout is considered related to user to access point ISP issues, or to a backbone ISP incident based on the number of users affected.
7. The non-transitory computer-readable medium of claim 1 , wherein baseline metrics are computed periodically, where the periods may be daily, weekly, or any longer or shorter time period.
8. A method implemented external to a network edge and located external from one of a computer device and a mobile device associated with a user, the method comprising steps of:
receiving, at a cloud-based system, telemetry data from a plurality of connector applications associated with a plurality of user devices, wherein the telemetry data is collected at the plurality of connector applications, the telemetry data being relative to performance of all tiers of Internet Service Providers (ISPs) from a point of view of the plurality of user devices;
generating baselines of the performance of the ISPs based on a plurality of metrics and the collected telemetry data, wherein the baselines are generated over a plurality of geo granularities;
training a Machine Learning (ML) model to assess blackout and brownout prediction accuracy at different performance values for the metrics; and
identifying a blackout or brownout and determining a severity of the blackout or brownout, wherein the blackout or brownout is identified when real time performance is worse than the performance baselines identified by the model, and wherein the severity of the blackout or brownout is based on a number of users affected and a geographic size of an affected area.
9. The method of claim 8 , wherein blackouts and brownouts are detected by computing distance of current metrics from baselines.
10. The method of claim 8 , wherein the collecting telemetry data comprises performing traceroutes from the plurality of user devices to frequently visited destinations.
11. The method of claim 10 , wherein the traceroutes are performed between data centers set up across the geographic locations.
12. The method of claim 8 , wherein the plurality of metrics comprise Page Fetch Time (PFT), Time To First Byte (TTFB), probe error rates, latency, packet drops, and route stability.
13. The method of claim 8 , wherein a blackout or brownout is considered related to user to access point ISP issues, or to a backbone ISP incident based on the number of users affected.
14. The method of claim 8 , wherein baseline metrics are computed periodically, where the periods may be daily, weekly, or any longer or shorter time period.
15. A method implemented external to a network edge and located external from one of a computer device and a mobile device associated with a user, the method comprising steps of:
receiving, at a cloud-based system, telemetry data from a plurality of connector applications associated with a plurality of user devices, wherein the telemetry data is collected at the plurality of connector applications, the telemetry data being relative to performance of all tiers of Internet Service Providers (ISPs) from a point of view of the plurality of user devices;
generating baselines of the performance of the ISPs based on a plurality of metrics and the collected telemetry data, wherein the baselines are generated over a plurality of geo granularities; and
training a Machine Learning (ML) model to assess blackout and brownout prediction accuracy at different performance values for the metrics, wherein a magnitude of a blackout or brownout is calculated as a function of the telemetry data and external data, wherein the external data includes social media and search engine keyword trends, and wherein the telemetry data is a function of Page Fetch Time (PFT), Time To First Byte (TTFB), and a score assigned to users of the plurality of user devices.
16. The method of claim 15 , wherein the collecting telemetry data comprises performing traceroutes from the plurality of user devices to frequently visited destinations.
17. The method of claim 15 , wherein the plurality of metrics comprise Page Fetch Time (PFT), Time To First Byte (TTFB), probe error rates, latency, packet drops, and route stability.
18. The method of claim 15 , wherein the magnitude of the blackout or brownout is represented in graphical form.
19. The method of claim 15 , wherein a blackout or brownout is considered related to user to access point ISP issues, or to a backbone ISP incident based on the number of users affected.
20. The method of claim 15 , wherein baseline metrics are computed periodically, where the periods may be daily, weekly, or any longer or shorter time period.