|
MDCSIM: A method and a tool to
identify services 1Universidade
Federal do Rio de Janeiro, Brazil
2CNPq fellow researchers {paulo.f.pires, fdelicato}@gmail.com Abstract Service identification is one of the biggest challenges in implementing a service-oriented architecture. Current service identification methods rely on business process descriptions to elicit the business perspective. However, service identification requires a level of business process documentation only found in organizations mature on business process modeling. Besides, current service identification methods have several drawbacks such as the lack of: (i) analyzing both business and IT domains, (ii) identification of both business and software services, (iii) service quality assessment; and (iv) method configurability. In this context, our work overcomes the aforementioned drawbacks by proposing a configurable service identification method (named MDCSIM) that uses master data, logical data models (obtained from organizations databases) and artifact-centric modeling technique. Master data (core enterprise information concepts, needed across different business processes, organizational units and applications across the enterprise) can be used as alternative input to business process. The logical data models aid the identification of master data attributes and contributes to the elicitation of IT perspective and identification of software services. Artifact-centric modeling technique is used along with master data to elicit business perspective and identify business services. MDCSIM also uses some metrics to assess service quality attributes in order to improve the quality of the identified services and of the SIM itself. MDCSIM is supported by a tool named MDCSIM plug-in. Such tool was implemented based on the Model-driven architecture (MDA). MDA enables the transformation of data logical models and artifact-centric models into models that describe candidate services. Finally, an initial assessment of MDCSIM is provided by comparing the service portfolios identified by using MDCSIM and by using other two data-focused service identification methods. Keywords: Artifact-centric modeling, Logical
data model, Master data, Model-driven architecture (MDA), Service
identification method (SIM), Service-Oriented Architecture (SOA). +: Corresponding author: Satoshi Tanaka IT Convergence Practice (INPRA), Vol. 2, No. 4, pp. 1-27, December 2014 [pdf] |