Analysis On The Structure Characteristics Of Software Network Based On Functions Invoking Relationships

In view of the stability and security of the software,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,path and community.Firstly,in order to demonstrate the information during the process of the software operation clearly,based on the invoking relationships between the function nodes,a novel method is proposed to construct the software network.By executing software several times in different system parameters,the information for the invoking relationships of the function nodes is traced.Then the execution results are handled and the model of software execution network is established.Secondly,considering the network software in dynamic execution process between the nodes of the calling sequence and the number of calls,based on this call.Analysis neighbor nodes of the direct and indirect contribution ability,through the algorithm CIWN(calInitialWeightOfNode)and CFWN(calFinal WeightOfNode)calculated node initial weights NIW(node initial weight)and the final weight NFW(node final weight).And through the algorithm MTKN(MinTopKNodes)to sort the weights of nodes,mining the key nodes in the network software.The distribution of key nodes in each version of the software is analyzed,and the characteristics of the key nodes in the process of software evolution are revealed.Again,in view of the complexity of the software function calls,and puts forward a based on key nodes representing the node path of the percentage of path mining algorithm IIEP(idenImporExePath)mining the critical path in the network,an important path to different versions of software analysis and guidance to software maintenance and prediction.Finally,introducing community structure in the software network,based on the key nodes,the algorithm COC(createOriginalCommunity)is demonstrated to initialize the community structure of the software network.Based on the dependence relationship,the nodes are divided into different communities and the optimal divided result can be obtained by the algorithm ETC(expandTheCommunity).Taking into consideration the characteristic of “high coupling and low cohesion”,the standard Q is used to evaluate the community structure after partitioning in the software network.With experiments for three 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.