Сохранить в:
Библиографические подробности
Главный автор: Manuja Bandal
Формат: Recurso digital
Язык:
Опубликовано: Zenodo 2022
Предметы:
Online-ссылка:https://doi.org/10.5281/zenodo.15044761
Метки: Добавить метку
Нет меток, Требуется 1-ая метка записи!
Оглавление:
  • <p><span>Testing software in cloud environments, especially within AWS infrastructures, presents distinct obstacles due to dynamic resource allocation, decentralized architectures, and unpredictable execution conditions. Conventional testing methods often fail to detect cloud-specific faults such as transient errors, race conditions in autoscaling, and inconsistencies in distributed systems. This paper introduces AI-Enabled Adaptive Fault Injection (AIAFI), a novel autonomous testing framework specifically designed for AWS-based applications. AIAFI autonomously detects, injects, and modifies fault scenarios in cloud-native applications through reinforcement learning (RL) and evolutionary search mechanisms. By utilizing AWS-integrated observability tools such as CloudWatch, X-Ray, and AWS Fault Injection Simulator (FIS), our approach enhances fault detection by 65% compared to leading-edge testing methodologies. Experimental results demonstrate that AIAFI effectively optimizes test execution, minimizes downtime, and improves the resilience of AWS-powered infrastructures.</span></p>