AI-DRIVEN PROGNOSTIC FAILURE ANALYSIS FOR AUTONOMOUS RESILIENCE IN CLOUD DATA CENTERS

The cloud data centers are considered the critical infrastructure for a wide variety of applications and services.Nevertheless, high availability and low downtime are still difficult to guarantee because of hardware failures, software anomalies and network disruptions.Traditional failure detection means, namely threshold based monitoring and rule based anomaly detection, lack in making predictions timely and accurate.An emerging solution to take such a proactive view to failure prediction and maintenance optimization is Predictive Failure Analysis (PFA) with machine learning (ML) power.Some recent advances of AI-driven failure prediction include the deep learning models such as Long Short Term Memory (LSTM) network, Transformer based architectures and Graph Neural Network (GNN) that can effectively learn failure patterns from large amount of system logs and telemetry data.Artificial Intelligence for IT Operations (AIOps) automates detect, understand, repair of failures which in turn reduces human intervention.Together, these innovations will make it possible to perform predictive maintenance, proactive fault management, and automated recovery, resulting in lowered operational costs and better reliability of the cloud.In this work, we study the state of the art of Omkar Reddy Polu https://iaeme.