摘要

A Dynamic quantitative analysis model for network survivability (DQAMNS) is proposed to measure survivability of large-scale distributed information systems based on ANP. Starting with the quantification and normalization of survivability factors (Resistance, recognition, recovery and adaptation, R(3)A), DQAMNS makes use of probability model and dimensionless representation to access the degree of each factor. Then, DQAMNS applies ANP to measure the weights of survivability factors based on experts' experience and gets a number between 0 and 1 which synthetically indicates network survivability. Analysis results show that DQAMNS can actually evaluate the four survivability properties, and precisely access survivability in different environments, thus they could be theoretical instructions to improve survivability in designing and implementing LSDIS.