A Graceful Degradation Model For Software Design And Implementation Based On Trustworthiness Measurement

As an important component of modern technology,the software plays an increasingly critical role in our daily lives and safety critical area.Once software fails,the resulting losses can be difficult to estimate.To build a trustworthy software and reduce losses caused by software faults,which is essential to improve the trustworthiness of software.This paper proposes a graceful degradation model based on trustworthiness measurement method.By establishing a hierarchical trustworthiness measurement model for the software system to measure and compute the software system’s trustworthiness in real-time,which can be achieved to detect and locate faults.Based on this,a version rollback algorithm for faulty components and faulty components selection algorithm for fault handling are proposed to improve the system’s fault tolerance and ensure the software system’s trustworthiness.This paper mainly focuses on the following three aspects:First of all,this paper proposes a hierarchical trustworthiness measurement model for software.Based on the system’s business logic,a hierarchical model for the software system is proposed,and a four-layer directed graph model,which can abstracte the software system as a directed graph with four layers is constructed by incorporating the dependencies among independent modules in each layer.While measuring the trustworthiness of the bottom-layer components,a comprehensive evaluation system for component trustworthiness has been established,which decomposes the trustworthiness of components into multiple trust attributes and calculates the trustworthiness of each component through a trust-based evidence model.To address the issue that different nodes in the directed graph model have varying impacts on the overall trustworthiness,a node importance model is proposed to distinguish critical nodes by combining the factors of node self importance and fault propagation capability.Subsequently,a hierarchical trustworthiness computing method is introduced to sequentially calculate the trustworthiness of each layer of the software system,leading to the ultimate calculation of the software system’s trustworthiness and trust level.Secondly,this paper proposes a graceful degradation model for software based on trustworthiness measurement.According to the definition of graceful degradation,and in response to the decrease in system trustworthiness caused by code vulnerabilities introduced by component secondary development,a fault localization method based on the software hierarchical trustworthiness measurement model is proposed to achieve fault detection and localization by real-time monitoring of trustworthiness changes in each layer of the system.Then,we proposes a software degradation model based on component version rollback to ensure system trustworthiness.In order to address the issue of component compatibility decline caused by version rollback,we proposes a component rollback version search algorithm based on binary search to search for the optimal rollback version of the fault component quickly.Furthermore,we proposes a fault component selection algorithm based on branch and bound method to search for the optimal selection scheme of fault components that meets the trustworthiness requirements of the system.Finally,this paper develops a hierarchical trustworthiness measurement and graceful degradation tool for software systems.This tool is based on the related calculation methods of the hierarchical trustworthiness computing model and graceful degradation model,and realizes the hierarchical trustworthiness computing function of the software system and visualizes the reliability measurement results of each layer,and realizes the automatic location of software faults and provides corresponding graceful degradation solutions and other functions.The tool combines theory with practice,and completes the entire process of building a software hierarchical trustworthiness measurement and graceful degradation model and implementing the corresponding tool,which has lowered the threshold for using the method and provided assistance for users to apply this method.