Research On Survivability Of Network System

As the development strategy of informationization driving industrialization is determined, the constructions of network and information system have got quite great progress. Network system is widely used in industry; business, government and the department of defend. Because the inside and outside environment of network system becomes more and more complicated, it is difficult to construct absolutely safe network system. Moreover, people increasingly depend on network, and more and more systems use distributed architecture, all of which bring great challenge to the network security technology. Survivability of network system is issued at this phase. Survivability aims to offer service and repair the corrupted service within a certain time despite the important part of system is damaged or destroyed, which makes sure the system be able to accomplish tasks in time.This paper studies mainly survivability theory and key technique of Network system, and introduces architectures and design methods of survivable system. In the paper, we specify survivable evaluation methods of system. The important works are as followed.1. Introduce survivable study background and significance of Network system, present research aspects and results of survivability done by some organizations or individuals in Domestic and Abroad.2. Through existing survivable evaluation methods, we analyze mainly combination method,modeling checking method and the state-based stochastic model method. And compare their merit and weakness respectively.3. Introduce theoretic information of Stochastic Petri Net, and depict its material application and analysis methods as a modeling tool.4. Propose a method on telecommunication switching system survivability analysis. Based on analyzing existing methods and defining the system attributes about survivability evaluation, the paper adopts Stochastic Petri Net (SPN) to analyze the survivability characteristics of the telecommunication switching system. The performance and fault-repair models are depicted by SPN and the quantitative approach of main survivability attributes are presented in the paper. At last, the simulation results show that the approach is correct and efficient.5. Present an approach by using Stochastic Petri Net (SPN) to quantitatively valuate two intrusion-tolerant server architectures. We introduce two intrusion-tolerant server architectures and define some availability measures of two architectures. Stochastic models are established by using SPN, and then we use assumed input parameter values in simulation experiments to obtain results, which present comparison of their survivability between the two architectures and show the relative strengths and weaknesses of the representative architectures. The results help the system designers to make informed choices for building more survivable service system.