Ben Amara Ayoub2025-02-092025-02-092024http://dspace.univ-khenchela.dz:4000/handle/123456789/7803Cloud computing is a paradigm that allows users to access shared computing resources over the internet, enabling rapid provisioning and scaling of resources with minimal management effort or service provider interaction. This research explores an innovative method for improving fault tolerance in cloud computing environments through intelligent virtual machine (VM) replication. The study specifically addresses the challenge of reacting to faults in hosts. It proposes a novel approach that considers cost factors and historical fault occurrences on specific hosts to intelligently select VMs for replication. The thesis demonstrates the effectiveness of this approach in improving fault tolerance through theoretical analysis, empirical evaluation, and practical implementation. Key findings include identifying fault-prone hosts, selecting VMs, and implementing proactive fault management strategies. It provides valuable insights and recommendations for cloud service providers and end-users, highlighting the potential benefits and challenges of intelligent VM replication.enThesis