Quantum Kernel Methods for Job Failure Prediction in Distributed Cloud Clusters

This project proposes the application of quantum machine learning (QML), specifically quantum kernel methods, to the problem of job failure prediction in large-scale distribut...ted cloud computing environments. Job failure prediction is a critical challenge for modern cloud schedulers because failures waste compute cycles and energy while slowing service delivery. Classical machine learning (ML) methods have demonstrated strong predictive abilities, but scalability and learning efficiency are being challenged by the increased complexity and dimensionality of modern datacentre workloads. This research will apply recent advances in quantum computing, specifically the use of quantum kernel algorithms, to evaluate the potential enhancements these models offer in terms of efficiency, accuracy, and generalization when applied to established failure prediction models. A curated portion of a publicly available cloud workload traces will be considered to train and test the quantum and classical models respectively using the job metadata, resource usage statistics, and task outcome information. Both quantum and classical models will be trained and compared using standard supervised learnings metrics.