大规模软件宏观拓扑结构的核数分析

Coreness Analysis on Large-Scale Software Macro-topology

针对大规模软件宏观拓扑结构体现出的层次特征,基于大量开源软件结构核数的统计数据,分析了节点核数的分布及与度值的相关性、最高核节点对各层的影响力,以及软件结构核数的演化趋势.分析结果表明,软件结构的核数普遍不大,节点核数分布具有无尺度特征,最高核节点影响力巨大;低度值节点的核数与其度值具有一定的正相关性,但是核数达到一定值后不再随节点度值增大,软件演化中其核数以及核心框架保持稳定.

For the macro-topology which shows hierarchical characteristics in large-scale software,coreness distribution and the correlation with degree,the effect of nodes with highest coreness on each and every layer and the trend of software structure coreness evolution analyzed on the basis of a large number of statistics data on coreness of open source software architecture.The result reveals that the k-coremax of software structure is generally small,the coreness distribution is scale-free characteristic,and the nodes with highest coreness have powerful influence. There is a certain positive correlation between coreness and degree for low-degree nodes, however, the coreness does not increase with the growing degree when the coreness comes up to a certain value, and the k-coremax and software core architecture remain stable during software evolution.