Analysis For The Structure Characteristics Of Software Execution Network Based On The Fault Propagation Probability

In recent years,With the stability problem and the safety problem of software system is becoming proniment.In this paper,based on the invoking relationship between nodes in the software network,introducing complex network theory,the structure characteristics of the software system is analyzed from the point of the network topology,in terms of node and community.Firstly,in order to obtain the specific information during the process of the software operation,this paper presents a modeling method which use variable size to describe the structure of software system.The method which based on the invoking relationship between function nodes,executing software several times in different system parameters,to traced the information of the invoking relationships of the function nodes.Then the execution results are handled by taking out redundant information and the model of software execution network is established.Secondly,taking into account the fault information can be transmitted between the function nodes in the software networks,the fault probability is defined as the vulnerability of the function node NMV(Node Measurement Value).Analyzing the invoking relationships during the software execution process,an algorithm ETN(Evaluate the Node)is provided to calculate the NMV value of a function node.In the end,the important function node is obtained.Thirdly,based on the label propagation community detection algorithm and the fault propagation probability between the function nodes during the software execution,a algorithm LCD(Label Community Detection)is provided to detect the community in software network.Which,use the sorting nodes by ETN to optimize the process of propogate the label.The detected communities can be used to analyse the software network and mining the potential characteristics.Finally,the two algorithms which proposed in this paper,is used to mining the important function nodes and communities on two types of open-source software.By analyze the important nodes and communities which is mining by the two algorithms in different version software,we can info the relevant information about the software system revolution law.With experiments for two types of software,the network structure of the software is analyzed,the potential characteristics of the software system are mined and the effectiveness for the proposed methods are verified.