Fault Tolerance (FT) is one of the cloud’s very critical problems for providing security assistance. Due to the diverse service architecture, detailed architectures & multiple interrelationships that occur in the cloud, implementation is complicated. A few other previous studies attempt to integrate the different fault tolerance frameworks and solutions suggested for the cloud environment, but some accounts occur to be constrained. This research article provides a detailed survey of the state-of-the-art research on emerging methods of fault tolerance for Cloud Computing & categorizes techniques of fault tolerance into three categories: Reactive Methods, Proactive Methods & Resilient Methods. Reactive Methods allow the system to reach a defect condition but instead attempt to get the device back up. Proactive Methods help to prevent the device from reaching a defective condition by introducing actions to minimize defects before impacting the device. On either side, newly developed Resilient Methods strive to reduce the amount of time a device takes to find from a malfunction. In the Resilient Methods context, Machine Learning and Artificial Intelligence played an important role in mapping the recovery period to a task to be configured. This survey offers a comprehensive and detailed description of the various faults kinds, factors, & different methods to fault tolerance used in the cloud. In light of their Basic Methods & some other specific characteristics, & also offers a Clear study of different tolerance mechanisms for failures & provides a comparative study of the structures under the article. It is noted that approaches of fault tolerance directed to checkpoint restart and replication are mainly included to address the crash faults in the cloud.

容错（FT）是云提供安全帮助的非常关键的问题之一。由于云中出现了多种服务架构，详细的架构和多个相互关系，因此实现起来很复杂。先前的其他一些研究尝试将针对云环境建议的不同容错框架和解决方案进行集成，但是有些帐户受到了限制。本文对云计算容错的最新方法进行了最新研究，并将容错技术分为三类：反应性方法，主动性方法和弹性方法。反应性方法使系统达到故障状态，但尝试恢复设备。积极主动的方法通过采取措施在影响设备之前最大程度地减少缺陷，来帮助防止设备出现故障。在任何一方，新开发的弹性方法都致力于减少设备从故障中发现所需的时间。在弹性方法上下文中，机器学习和人工智能在将恢复周期映射到要配置的任务中发挥了重要作用。该调查对云中使用的各种故障类型，因素和不同的容错方法提供了全面而详细的描述。根据它们的基本方法和其他一些特定的特征，还提供了对故障的不同容忍机制的清晰研究，并对本文下的结构进行了比较研究。