MUMBAI, India, Sept. 7 -- Intellectual Property India has published a patent application (202641104454 A) filed by Ms. Dhanalaxmi Chinthala on August 31, 2026, for System And Method For Causally Verified Anomaly Detection And Response In Distributed Computing Networks.
Inventors include Dr. Lingam Sunitha; Ms. Dhanalaxmi Chinthala; Mr. Vinayak; Ms. Swetha Polisetty; Ms. Shaikh Sumaiya; and Mr. Satheesh Kumar S.
The application for the patent was published on September 04, 2026, under issue no. 36/2026.
Abstract: ABSTRACT SYSTEM AND METHOD FOR CAUSALLY VERIFIED ANOMALY DETECTION AND RESPONSE IN DISTRIBUTED COMPUTING NETWORKS The present invention provides a method and system for detecting anomalies and taking responsive actions that have been causally verified in distributed computing networks. Multimodal telemetry including network traffic telemetry, performance telemetry, execution telemetry, application telemetry, security telemetry, workload telemetry, configuration telemetry, and communication link telemetry is collected and temporally synchronized. A time-varying dependency graph correlates the telemetry with computing nodes, services, workloads, data resources, and communication links. Machine learning algorithms predict a candidate anomaly, a potential source thereof, and a propagation signature. Anomalies and competing non-anomaly explanations are postulated. An intervention with bounded impact and expected information gain is selected. Post-intervention telemetry is compared to the predicted signature to evaluate whether the anomaly is causally consistent, to identify an anomaly-source resource, and to pinpoint a propagation path thereof. Remediation is approved only when a verification criterion has been met. Otherwise, remediation is prevented, delayed, or adjusted. Thus, incorrect and unnecessary disruption of the network is avoided.
Disclaimer: Curated by HT Syndication.
