Functional break down

The following graphic shows more details for a deeper understanding of our Airlock Anomaly Shield and Security Gate services:

ML Component Overview 2
  • Description:
  • Within an Airlock Gateway instance, the Security Gate service is the core service. It processes all requests and, based upon the enforcement logic, performs security actions such as blocking requests.
  • Only one aggregator process is active for every session. This avoids performance issues in the gateway and optimizes latency.

  • To this strong conventional security approach, Airlock Anomaly Shield adds dynamic machine learning security by computing session metrics into anomaly indicator values. Depending on the anomaly indicator values, actions can be triggered by the Security Gate.
  • During model training, the model parameters are being trained and refined. This is required for the subsequent machine learning model evaluation process.
  • Model evaluation is the Airlock Anomaly Shield core process that computes aggregated session data against trained model parameters in order to generate anomaly indicator values for the Security Gate service.