Anomaly Detection

Anomaly Detection Oveverview

Anomaly detection is based on the forecast of the ML models that analyze the profiling metrics. The models can predict the expected values and ranges of the metrics based on historical and current data. If the actual values of the metrics deviate significantly from the predicted values, an anomaly is detected.

Upon the detection of one or more anomalies, an incident is created. An incident is a record that contains information about the anomaly, such as the data source, the metric, the value, the timestamp, the severity, and the root cause. An incident also provides suggestions and recommendations on how to resolve or prevent the anomaly by providing access to the other data governance features of Blindata.

Software Requirements

SaaS Installation

Blindata Agent:

  • The SaaS installation requires the Blindata Agent to be installed and configured. This lightweight software component serves as the data collector, responsible for gathering profiling metrics from the customer’s network.

Access to Customer’s Network:

  • The Blindata Agent must be granted appropriate access permissions to the customer’s network. This access is essential for collecting real-time data and metrics required for profiling and anomaly detection processes.

On-Premises Installation

Blindata Agent:

  • Similar to the SaaS installation, the on-premises installation necessitates the installation and configuration of the Blindata Agent. This component is responsible for collecting profiling metrics directly from the customer’s on-premises infrastructure.

Blindata-ML Service:

  • In addition to the Blindata Agent, the on-premises installation requires the deployment of the Blindata-ML service. This service incorporates machine learning algorithms and models for advanced anomaly detection and predictive analysis. It processes collected data to identify anomalies and generate forecasts.

System Requirements:

  • Ensure that the hardware and software specifications for hosting the Blindata-ML Service meet the system requirements. This includes considerations for processing power, memory, storage, and any other dependencies.

By adhering to these software requirements, both SaaS and on-premises installations can ensure the effective deployment and operation of the Blindata profiling and anomaly detection system tailored to the specific needs of the customer.