Through comparing the relationship between current response time of key service request and confidence interval of its history average response time

this paper presented an autonomous configuration algorithm for improving network survivability（abbreviated as NSAC） based on central limit theorem and hypothesis testing theory.According to different durations in the process of key services execution

the algorithm dynamically dealt with two instances for taking the timely policies such as resource reconfiguration

service degradation or service roaming

etc.As a result

key services could be accomplished farthest within its deadline expected by users.Moreover

the whole configuration process is transparent for users.Simulation experiments show that the proposed method can improve the survivability of key service effectively under the circumstance of the network light load and overload.