Abstract

The SAVI approach to integration embodies three fundamental concepts: (1) an architecture-centric emphasis (wrapped around an annotated architectural model with analyses carried out at the system level after modifications); (2) a component-based decomposition of elements of the system that support a building block approach; and (3) a unique level of consistency checking to assure compatibility for the physical and logical integration through ana "single truth" across the multi-domain model set. Thus, this unique architectural model implements the SAVI mantra of "integrate, analyze -- then build" to address a range of virtual integration issues. Careful safeguards to protect the integrity of intellectual property for each member of the development team are provided through the SAVI Model Repository and Data Exchange Layer (SMR/DEL). These two core information-sharing elements of the VIP will eventually depend heavily on standards-based (likely ISO 10303-239) information exchange and the SAVI team has taken steps recently to cooperate with global collaborators, both in the United States and in Europe in this type of information exchange. This sort of standards-based data exchange also offers considerable promise for protection of sensitive information within a system development with competing suppliers. The paper describes how multiple architectural definition languages (specifically SysML and AADL) have been utilized in developing the core of this model-based analysis approach. The objective is to exploit strengths of both these languages while maintaining capability to translate between both variations of architectural models. The primary means of accomplishing this two-way translation is an extension of the translator generated by Cofer, et al, for the DARPA META program. This translator currently operates in a two-way sense (between SysML and AADL) only upon a rather limited subset of SysML capabilities but the SAVI team expects to see this two-way capability broadened with each incremental development phase of the VIP. Extensive use of recent new annexes for AADL, notably the Error Model Annex, has been a means of automating system safety analysis tools (like Functional Hazard Assessments, Failure Modes and Effects Analyses, and Fault Tree Analyses) that underpin broader System Safety Analysis. Another characteristic embedded in the SAVI VIP is a comprehensive form of consistency checking designed to evaluate integration of components and alterations of such components. At least six types of consistency are considered in a SAVI-compliant integration effort: (1) interface consistency, (2) compositional consistency, (3) constraint consistency, (4) behavioral consistency, (5) version consistency, and (6) verification consistency. Moreover, whenever a modification is made to any component, not only are these elements of consistency to be addressed, the effect of the modification on the entire system must be quantitatively analyzed using the SAVI annotated architectural model for the modified system. Each system considered under the SAVI paradigm is analyzed for system properties required to meet its performance specification using this architectural approach. The AADL portion of the architectural model structure allows quantitative evaluation of the system impact of each change made to the substructure. In this sense, the SAVI approach facilitates quantitative trade studies aimed at the system behavior during each iteration in the design loop.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call