XAIOps detects the real-time load status and abnormal patterns of the operating system through data collection and learning on various
IT infrastructure environments, including applications/databases/servers/networks/unstructured logs, etc. Moreover, it is an AI-based IT operation
intelligence solution that supports the whole IT operation. By providing proactive responses, XAIOps will help users to predict future failure situations.
Deep Learning-based Solution Optimized for Time Series
Pro-Active Response to Failure with Intelligent Future Prediction
Integrated Anomaly Cause Analysis
for Various IT Resources
Verified by Large Financial/Manufacturing/Distributing Clients
XAIOps is built on machine/deep learning methodology, which enables real-time data collection and pattern learning for intelligent IT operations. This includes more accurate anomaly detection, load prediction, root cause analysis, and proactive alarm provisioning, all of which are integral to XAIOps' DNA.
Through AI automated diagnosis, XAIOps provides a real-time integrated monitoring dashboard where users can check the entire IT system environment, including application services, instances such as WAS/DB, and infrastructure such as servers/networks.
Based on the AI anomaly detection model, XAIOps determines whether there is an anomaly in each system area (Service/Instance) in real time and provides an alarm according to the anomaly judgment level. Real-time monitoring results are offered visually as well.
AI Monitoring allows users to detect real-time failure based on collected IT operation data. When a failure occurs, XAIOps provides root cause of failure and detailed connection information through automatic inference/analysis.
A confidence interval (Base-line) is generated through past data learning, and when real-time observations fall outside the range, it will be detected as an anomaly. Proactive response is possible by predicting the situation 30 minutes to 1 hour into the future.
XAIOps detects abnormal log patterns in real time by learning on various types of unstructured log files (including Biz Log, Sys Log, Was Log, etc.). What’s more, it is possible to check log messages and analyze correlations.
Short-term load prediction is an analytic function for forecasting key performance indicators in the short term: within one week into the future.
Long-term load prediction is designed for analyzing long-term key performance indicator forecasts from 1 to 12 months into the future and is mainly used for transaction volume and expected usage of system resources.
Performance Statistics is useful for analyzing anomaly detection, predictive model performance by setting a desired analysis period, or analyzing past pattern types. You can confirm the performance (accuracy) of your operating model and use it as a baseline for activities’ improvement.
When an instance failure of a specific server occurs, it is possible to prevent the expansion of failure in advance by analyzing in detail which indicators were caused and which systems and problems may occur in related calls due to the failure of the instance.
In the event of a specific application service failure, XAIOps provides trend analysis of key performance indicators that affected the failure and trace analysis of performance delay intervals and causes. Detailed connection tracking is also easily checked by linking call analysis between services.
With XAIOps, users can easily check the status of events by date and statistics by grade for the past week and analyze the types of events that have occurred based on the date of failure in detail.
Establishment of an AI-based Intelligent Integrated Monitoring System
Efficient and stable system operation is required by adopting the latest AI-based technology to the core business.
• In the event of an IT failure, rapid failure detection and prediction provides us with proactive response to failure.
• With AIOps, analysis time of the cause of failure has been shortened to within 5 minutes after establishing a rapid
• Obtaining integrated analysis and data utilization plan by collecting various IT operation data.
• Improved service satisfaction of employees through stable operation and pro-active response of the core business
Establishment of an AI-based Intelligent Detection/Prediction System
There was a need to build an intelligent failure prediction system for proactive failure response.
• Verification of anomaly detection and predictive system possibility by establishing various IT operation data
• Establishment of a proactive failure response system through rapid failure detection and confirmation.
• Reduction and clarification of communication between departments as a comprehensive analysis of causes of
failure becomes possible.
• With intelligent control, work concentration has been improved through efficient redeployment of IT operation personnel.
Establishment of an Intelligent ICT Integrated Monitoring System
There was a need to build an intelligent control system through enterprise-wide monitoring data integration.
• Establishment of centralized performance and failure analysis response system through collection of various IT
• With application of the latest Machine Learning/Deep Learning-based algorithms, prompt ICT operation control
becomes possible through improving the accuracy of failure detection and cause analysis.
• Establish a proactive failure response system through rapid failure detection and prediction.
• Clarification of failure cause analysis, better and more efficient communication between development / operation