The high cost of changes is a primary obstacle to more frequent software upgrades. We suggest addressing this problem by introducing an automated framework and methodology, and a mix of technologies to identify the impact of changes and upgrades (due to error fixing and functionality enhancement) and from component replacement within a single product and a product family. This methodology will improve the reliability of networked software and reduce validation costs by implementing an innovative solution for the automatic detection and localization of program bugs.We note that we do not address the initial analysis and validation, but aim to reduce the costs of testing changes in existing systems by several orders of magnitude by introducing automatic techniques to analyze and test software increments. The goal is to make analyzing and testing networked software after each change practical, which allows developers and testers to diagnose and remove faults as soon as they occur, thus saving testing and development effort while increasing software reliability. The new analysis and testing techniques that we suggest are created by improving state-of-the-art techniques that apply both static and dynamic analysis. Stand-alone solutions for verification of system changes do not scale for complex environments, and any new industrially applicable solution must integrate several technologies into a common framework. Moreover, as we discuss in more detail later, some program constructs pose a challenge for one technology while being easy or at least feasible to solve using another technology or a combination of technologies.We suggest a combination of techniques that creates an integrated paradigm for the analysis of software upgrades. The key approach in our methodology is to automatically verify that safety properties that hold for a given version continue to hold after a change without repeating the entire analysis. This is achieved by means of analysis procedures that focus on the parts of the system that have changed during evolution of a design or are affected by this evolution, and determine whether all behaviors of the original system are preserved in the new version of the design. The analysis is performed by the integration of technologies that apply different types of analysis (static and dynamic) andwork at different abstraction levels. These technologies address different classes of problems and share a common solution framework. Clearly, every technology has its limitations—some of the limitations are inherent to the technology itself or to a particular component of the technology; others are challenges that can be dealt with by combining the technology with others. In what follows, we discuss the challenges in more detail.
Chockler, H., Kroening, D., Mariani, L., Sharygina, N. (2015). Challenges of existing technology. In Validation of Evolving Software (pp. 7-17). Springer International Publishing [10.1007/978-3-319-10623-6_2].
Challenges of existing technology
MARIANI, LEONARDOPenultimo
;
2015
Abstract
The high cost of changes is a primary obstacle to more frequent software upgrades. We suggest addressing this problem by introducing an automated framework and methodology, and a mix of technologies to identify the impact of changes and upgrades (due to error fixing and functionality enhancement) and from component replacement within a single product and a product family. This methodology will improve the reliability of networked software and reduce validation costs by implementing an innovative solution for the automatic detection and localization of program bugs.We note that we do not address the initial analysis and validation, but aim to reduce the costs of testing changes in existing systems by several orders of magnitude by introducing automatic techniques to analyze and test software increments. The goal is to make analyzing and testing networked software after each change practical, which allows developers and testers to diagnose and remove faults as soon as they occur, thus saving testing and development effort while increasing software reliability. The new analysis and testing techniques that we suggest are created by improving state-of-the-art techniques that apply both static and dynamic analysis. Stand-alone solutions for verification of system changes do not scale for complex environments, and any new industrially applicable solution must integrate several technologies into a common framework. Moreover, as we discuss in more detail later, some program constructs pose a challenge for one technology while being easy or at least feasible to solve using another technology or a combination of technologies.We suggest a combination of techniques that creates an integrated paradigm for the analysis of software upgrades. The key approach in our methodology is to automatically verify that safety properties that hold for a given version continue to hold after a change without repeating the entire analysis. This is achieved by means of analysis procedures that focus on the parts of the system that have changed during evolution of a design or are affected by this evolution, and determine whether all behaviors of the original system are preserved in the new version of the design. The analysis is performed by the integration of technologies that apply different types of analysis (static and dynamic) andwork at different abstraction levels. These technologies address different classes of problems and share a common solution framework. Clearly, every technology has its limitations—some of the limitations are inherent to the technology itself or to a particular component of the technology; others are challenges that can be dealt with by combining the technology with others. In what follows, we discuss the challenges in more detail.
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
