Based Services To Multi-state Survival Model Of The Ip Network And Shift Algorithm

Network security is always the key issue of computer network and its application. Traditional theory and technology of network security have three insuperable flaws. First, centralized control method is incompetent to distributed network environment; second, isomorphs of network structure can't prevent the rapid and extensive spread of suspicious intrusion and virus; final, because the thread and technology of security can't adapt to the changeful network environment. The research of network survivability is a breakthrough and innovation to the traditional network security concept, emphasizes the ability of completing the major mission timely even when the network service system suffers attacks, failures or accidents. The key idea of survivability is the system can complete the mission and repair the injured service timely when the intrusion occurs, even the vital part of system damages or is destroyed.According to the current analysis and discussion of the correlative survivability technologies, this paper presents a layered survivability model of diversity design. This idea originates from the favorable adaptive capability of gene diversity, biology diversity and gene mutation to the uncertain changes of environment, its main viewpoint is to design multi- implements of different bugs, but same function for every network service, and then improve survivability ability of service system through shifting from different implements to transferring bugs of every implement itself. Based on this layered survivability model, this paper does summary designs for service request registration and communication interface of same layer, and carries through a feasibility analysis and functionality verification by using a Markov state transferring model. In the precondition of above diversity design system, this paper also puts forward a random tenure arithmetic of different implements'competition. This arithmetic randomizes different implements'running time, shifts actively to increase attackers'detecting time to system bugs, and optimizes service operation of different implements through statistic of tenure time.