Abordagem baseada em Cenários para Extrair Requisitos de Tolerância a Falhas de Sistemas-de-Sistemas a partir de Processos-de-Processos de Negócio

Abstract

Context: Systems-of-Systems (SoS) refer to a grouping of software systems, called constituent systems, from different organizations with their specific business processes. When such processes are put together to achieve strategic objectives of alliances of organizations (such as partnerships, mergers or acquisitions of companies), they give rise to large, complex and dynamic business processes. These singular processes are usually automated by SoS and are called Processes-of-Business Processes (PoP). From this perspective, to achieve the strategic objectives of the alliances, the constituent systems interoperate among them to obtain behaviors not provided by a single software system. Problem: Failures can occur during this interoperability affecting the behavior and reliability of the whole SoS. Despite the importance of fault tolerance requirements for the SoS stability, they are still not widely explored in the literature. Existing works generally consider these requirements only during architectural design, neglecting the importance of requirements management throughout the entire SoS lifecycle and the alignment between technical and business levels. Objective: The main objective of this master's work is to define an approach based on scenarios in BPMN (Business Process Model and Notation) notation to systematically extract SoS fault tolerance requirements related to communication problems between constituent systems from useful PoP information. This approach consists of a metamodel based on exception handling elements from BPMN, commonly used to represent communication failures between business processes of alliances of organizations; and a set of automatic guidelines for SoS fault tolerance requirements extraction based on elements from this metamodel. The exception handling elements of this metamodel were identified from abstract scenarios related to exception handling during the message sending and receiving. These scenarios were built based on acquired knowledge from the modeling of five concrete scenarios of real PoP from four distinct domains (i.e., educational, agribusiness, emergency, and hospital health). Research Method: The work was developed based on the Design Science Research by producing the following artifacts: two abstract scenarios, a metamodel, and a set of guidelines to extract SoS fault tolerance requirements. Results: The work particularly supports the extraction and specification phases of fault tolerance requirements during SoS Requirements Engineering from information obtained in PoP models. Thereby, it is possible to achieve the alignment between the technical and business levels from the perspective of fault tolerance during the communication among the constituent systems, which could contribute to the competitiveness and profitability of alliances of organizations.