Patent · US Active

Machine learning-based deployment of guardrail policies

US11349911B1 · kind B1 · utility

1Cited by
0References
20Claims
0Family size

Assignee

Inventors

Key dates

Filing dateApr 14, 2021
Grant dateMay 31, 2022
Priority date
Expiry dateApr 14, 2041

Classification

  • Technology area (CPC H)Electricity
  • CPC primaryH04L63/0227
  • WIPO fieldDigital communication
  • WIPO sectorElectrical engineering

Abstract

A system can receive a guardrail policy request that specifies a guardrail policy to assess for deployment on a server to protect at least a specific port of the server. The system can execute a fingerprint clustering machine learning model using server fingerprint data to generate cluster data that identifies a virtual machine cluster that includes a plurality of virtual machines executed by the server. The system can execute a traffic discovery machine learning model using server traffic data and the cluster data to generate a confidence score indicative of whether deployment of the guardrail policy would have an adverse impact on the server. The system can execute a risk assessment machine learning model using the application type data to generate a risk assessment score. The system can evaluate the confidence score and the risk assessment score and can determine whether the guardrail policy should be deployed on the server.

Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.