Cloud computing survivability is a new direction in network application security research. A system survivability emergency response strategy was constructed in the wide area network environment of cloud computing, which employed dynamic service drift technology and task scheduling method technology to achieve dynamic service drift in the case of a large number of service nodes. Dynamically distributing cloud services and data across multiple service nodes can eliminate the single failure point of traditional services. In order to enhance the triggering mechanism of service drift, an adaptive random autonomous scheduling algorithm was proposed, which compared with the traditional drift mechanism, added triggering conditions and introduced ant colony algorithm for scheduling. The conclusions show that the improved service drift model have significantly optimized anti-destruction ability, drift time, and drift efficiency, which can better ensure the continuity and reliability of the services provided, thus improving the survivability of big data service in the cloud computing environment.