Large-scale Business Support Systems.

  • For large-scale business support systems in the telecommunications industry, system abnormal events are usually rare events, that is, most of them are normal events, and the number of abnormal events is very small. How to predict the occurrence of system abnormal events is very challenging.
  • The introduction of 1H3A can successfully solve this key issue, extract important features from the original data, predict the upcoming system abnormal events in advance, and use root cause analysis technology to find the cause of the abnormalities, so as to eliminate the upcoming system abnormalities in advance and avoid system abnormal events.
  • Other Telecom Fields

    Mobile Billing System

  • It includes 50 hosts, 4000 check-in personnel, 3600 monitoring value indicators, 4320 log files are generated a day, and 700MB log volume is processed a day.
  • Through the currently available accounting system probe information, predict whether the system status will be abnormal in a short period of time (10~60 minutes) in the future.
  • Use past system history probe records to analyze the correlation between probes and find out important probes that significantly affect the system status.
  • Avoid losses caused by abnormal system status through important probes that affect system status and accurate system status prediction.
  • Internal Cloud System

  • Including 79 servers + 84 switches, a total of about 100,000 monitoring indicators.
  • Video and VPN system

  • Including mysql service on 67 virtual machines + 12 network devices, a total of about 80,000 monitoring indicators
  • MySQL DB quality monitoring: MySQL currently has a large number of open source DBs in the company. Each VM in the system has a MySQL DB. Stable service quality is the key.
  • Network quality monitoring: The network performance of the system is not good, and some users reported that the network connection is slow. According to the maintenance personnel's observations, the network-related indicators are all normal, and new solutions are needed to improve problem detection and assist in speeding up problem diagnosis when problems occur.
  • Cooperating with us, possible major system vulnerabilities can be prevented in advance.

    View application cases