一种基于进程代数的软件可信性度量模型

A Software Trustworthiness Metric Model Based on Process Algebra

基于软件可信本质和进程代数给出了软件可信性评估框架及度量指标。将软件正向和逆向工程都统一到进程代数这一尺度上,这样进程代数中进程的等价程度便可表示软件的可信程度,即根据软件执行踪迹进行逆向工程,恢复出表示软件行为的UML顺序图并转换为进程代数,将初始的高层软件需求(UML协作图和状态图)也转换成进程代数,通过分析所得到的两个进程,看其等价程度,等价程度越高则软件可信程度越高,不等价时计算软件可信性度量指标TD(trustworthiness degree),其数值越大,则软件可信性越高。

The essence of software trustworthiness is its dynamic behavioral trustworthiness. And the dynamic behavior of software can be described by using process algebra. Based on the essence of software trustworthiness and process algebra, a trusted evaluation framework and metric model about the trustworthiness of software were proposed in this paper. Forward and reverse engineering of software were represented by process algebra, so the degree of software trustworthiness can be denoted by the extent of the two process equivalence. That is, reverse engineering based on software execution trace was conducted. The recovered UML sequence diagram embodying software actual behavior was converted into the implementation - level process; and corresponding UML collaboration and state diagram reflecting software initial behavior was also transformed into the initial - level process. Through the comparison and analysis of such two processes, the consistent degree of the behavior of implementation - level process conforming to the behavior initial - level can be verified. The higher the level of their equivalence, the higher the degree of soft-ware high trustworthiness. If no equivalence exists in the two processes ,the TD （trustworthiness degree） of software will be evaluated. Similarly,the greater numerical value of TD is,the higher software trustworthiness is.