MUMBAI, India, Aug. 31 -- Intellectual Property India has published a patent application (202641101546 A) filed by B Subramanya Anil Kumar Research Scholar Department Of Cse; and Dr Basant Sah Associate Professor Department Of Cse on August 22, 2026, for A Hybrid Machine Learning-Based Framework For Real-Time Virtual Machine Failure Detection, Dynamic Task Rescheduling, And Qos-Aware Cloud Resource Management.
Inventors include B Subramanya Anil Kumar Research Scholar Department; and Dr Basant Sah Associate Professor Department Of Cse.
The application for the patent was published on August 28, 2026, under issue no. 35/2026.
Abstract: The proposed invention relates to cloud computing and, more particularly, to an intelligent framework to detect virtual machine failure, dynamic task rescheduling and Quality of Service-aware cloud resource management. The framework continuously collects cloud and virtual-machine performance parameters including CPU utilization, memory utilization, network traffic, energy consumption, task execution time, task priority and VM status. The collected data is preprocessed, and relevant features are extracted for machine-learning-based prediction. A hybrid prediction mechanism combining Light and Random Forest is configured to identify abnormal resource-utilization patterns and predict potential virtual-machine failures. Upon detecting a potential failure condition, tasks associated with the affected virtual machine are dynamically rescheduled to one or more suitable alternative virtual machines. The framework further evaluates QoS parameters including reliability, availability, execution time, energy consumption and SLA compliance and dynamically manages cloud resources according to the predicted conditions and QoS requirements. The invention thereby provides an integrated mechanism for proactive failure management, task continuity, resource optimization and QoS-aware cloud operation.
Disclaimer: Curated by HT Syndication.
