Wireless Sensor Network is one of fast feeder for emerging technologies like of internet of Things. These networks are established to operate at a different condition which includes non-human intervene sites. These sensor nodes range from specific purpose to multi-purpose, multi-home: accommodating more sensor under a single architecture. The cost of sensor and deployment of these under hostile situation makes them costlier network. In order to provide access to this sensor network on-demand, network virtualization is created on the cloud called as cloud-sensors. Cloud while providing a sensor as a service experience errors/faults. Here we survey for the appropriate methods that can handle fault and recovers system at every level including at physical sensor level, content delivery level and physical sensor to virtual sensor mapping level. Therefore, the research work attempts to explore efficient reliability models at each level and helps in building a comprehensive reliability model for Virtualization of Wireless Sensor Network.