The CuriX® Cycle

How we get to know your system and predict anomalies of any kind.

When we install CuriX® in your system, we first start the learning phase, where we get to know the normal state. In order to fully utilize the built-in artificial intelligence, each new installation goes through a four-step cycle after which we can predict any anomalies and problems in your system.

 

 

 

 

The 4 steps of the cycle

From Data to Information via Knowledge and Wisdom

 

CuriX Machine Learning Circle

Step 1: Data

Feed CuriX® with a wide variety of data sources

 

  • Unrestricted data sources
  • Dynamic data
  • Static data
  • Automatic aggregation
  • Proprietary and custom interfaces
  • Dynamic customizations

CuriX® can incorporate unrestricted dynamic and static (e.g. CMDB) data. In the event of intentional system changes, the data collection is dynamically adapted to the new circumstances. Any data sources can be integrated at the same time. CuriX® automatically aggregates across all these sources. The data can originate from log file management, SIEM, monitoring, IoT, cloud, etc. It does not matter if it is a custom or proprietary interface. If a real data lake is missing, we can also implement CuriX® in combination with a log file management system that is ideally suited to your needs.

Step 2: Information

CuriX® gets to know your system

 

Based on the collected data, CuriX® automatically learns the normal state of the system with all integrated data sources and defines this as the baseline. Deviations from the baseline are detected as anomalies. CuriX® uses artificial intelligence to determine whether an anomaly is critical. If necessary, further steps are taken, such as setting off an alarm. In addition to baseline-based detection, threshold values can also be defined. Compared to classical solutions, CuriX® allows a proactive prediction of threshold violations, which can thus be avoided in advance.

All integrated data are automatically correlated to derive dependencies. The combination of static and dynamic data forms a holistic model, which allows statements beyond the technical level to the entire organization. This can even be done across layers from the application to the infrastructure. In general, the more data CuriX® is provided with, the more and more detailed analyses are possible.

  • Learning the baseline
  • Automatic anomaly detection
  • Threshold based predictions
  • Automatic cross-layer correlation of all data
  • Linking of static and dynamic data
  • System model generation
  • Holistic model

Step 3: Knowledge

CuriX® understands the severity of errors

 

  • Determination of the health of overall and subsystems (health score)
  • Determination of critical deviations from normal behavior
  • Prediction of possible errors and system failures
  • Determination of possible weak points

Based on the detected anomalies and the correlations, critical system deviations are determined and predictions for possible future failures as well as system failures are made. Unknown patterns from inside and outside are detected, which significantly increases system security and resilience. All factors flow into a scoring system, which provides you with clear indications of critical conditions. The health status is indicated for the entire system as well as for all subsystems with a health score. This serves as a basis for root cause analyses, healing advices and self-healing.

Step 4: Wisdom

Your system becomes resilient.

 

In the CuriX® Dashboard, all important parameters are visible on one page and are easy to use. Because known correlations can be used to determine the cause of the problem, CuriX® provides you with guidance on how to fix the problem (Heal Advices). The operations team receives these messages early, leaving enough time to act. Issues can be resolved before they impact customers and the business. Forwarding of Predictions and Heal Advices via mail or into a ticket system is standard. A ticket can be forwarded directly to the person responsible for the problem. In self-healing mode, CuriX® can take semi- or fully automated measures depending on the problem at hand.

  • Root Cause Analysis
  • Determination of error locations
  • Support for problem solving through Heal Advices
  • Automated triggering of preventive measures (self-healing)
  • Dispatching / forwarding of tasks / tasks
  • Connection to external ticketing, alarm systems, etc.

CuriX® Features

Why CuriX®?

CuriX® is suitable for all kinds of system landscapes – on premise, hybrid, or cloud only – and can therefore be used flexibly. CuriX® is based on your data lake. This can be fed by SIEM, monitoring, or other log file collection tools. It is therefore suitable as an add on as well as a replacement for your monitoring tool – of course across data silos.

  • Fully automated correlation/detection of data in real time
  • Integration of metrics without configuration effort
  • Multiple approaches to anomaly/fault detection
  • Systemic failure prediction / Root cause error analysis
  • Root cause analysis
  • Intelligent alerting / alert forwarding
  • Healadvice / Selfhealing
  • Connection of Business Model and System Architecture (IT)
  • Integration of source-independent data (agnostic)
    • IoT, cloud, IT, smart devices, metrics, NDR, EDR
  • Integration of target systems (agnostic)
    • Risk management tool, decision support systems, ticket systems, SOAR, alert management
  • Cross-silo data analysis
    • Among others monitoring logfile and SIEM tools
  • Noisefiltering/-reduction