Network state model construction method and state evaluation method based on differential manifold

Abstract

The invention discloses a network state model construction method and a network state evaluation method based on a differential manifold. The invention can make the network security evaluation no longer depend on the traditional topology path, so that the calculation result has better integrity and calculation scalability. The invention abstracts the network system state as a high-dimensional differential manifold, and characterizes the network state in more detail. The network state corresponding to each point in the manifold is presented by all the attributes of each device node in the network system. The network state can be more objectively reflected by evaluating the attribute information of each device node in the network system. The network state can cover all the nodes in the network system and their attribute information, thus ensuring the integrity and comprehensiveness, and the calculation scalability is good. At the same time, the invention does not need to consider the path problem between nodes, and the path explosion problem does not exist.

Description

technical field [0001] The invention relates to the technical field of network security assessment, in particular to a method for constructing a network state model based on a differential manifold and a network state assessment method. Background technique [0002] Most of the network evaluation methods have been based on the topological path of the network and the evaluation method of the calculation of the probability on the path. For example, evaluation methods based on attack graphs, evaluation methods based on Bayesian networks, and so on. These methods abstract the network and establish a network topology model, and assign values ​​to each node on the model and the attributes on the paths between nodes. The assigned attributes will vary depending on the calculation method. Then, through a certain probability calculation method (such as Bayesian probability, etc.), calculate the risk probability value of a certain path or a certain network model being attacked, and us...

AI Technical Summary

Problems solved by technology

[0003] The scale of a real network system is often very large: there are a large number of nodes (hosts or service devices) in the network system, and the paths between nodes are intricate. It is very difficult for the existing evaluation methods to exhaust all possible paths, and it is easy to appear. Some paths are missing or it is difficult to exhaust all possible paths, which makes the final calculation result incomplete

At the same time, with the linear increase in the number of nodes in the network model, the growth trend of the path quantity in the entire network model is exponential, and the calculation amount will also increase exponentially, which greatly increases the difficulty of calculation. low efficiency

Images