Applying a Model-Driven Framework for Gap Analysis: Towards Business Service Engineering

Introduction

Service orientation utilizes services as building blocks to support rapid, low-cost and easy composition of distributed applications. The key to this concept is Service-Oriented Architecture (SOA), which is a logical way of creating loosely coupled, interoperable business services that can be easily shared within and between enterprises, via published and discoverable interfaces. Service-oriented development methodologies such as: Arsanjani et al. (2008), Brown et al. (2005), De Castro et al. (2009), De Castro et al. (2011) and Papazoglou and Van den Heuvel (2007); provide models, best practices, standards, and reference architectures needed to construct a well-defined SOA (Papazoglou, 2008).

Central to any service-oriented development methodology is business service analysis (Nayak et al., 2007) which aims at identifying, conceptualizing and rationalizing business processes as a set of interactions. Thus, starting from the business process models, analysts can identify candidate business services related to these processes. The concept of business service goes beyond the simple idea of service as understood in the context of SOA; it describes a business task relevant to a business domain (Nayak et al., 2007).

In this chapter we shall focus on the problem of how existing applications can be traced to new SOA environments that are more than basic wrapping options, tackling the problem from the standpoint of business service engineering, the main objective of which is the re-use of business process functionalities in new composite applications. To achieve this objective it is necessary to start by capturing the “existing” application portfolio and software resources, including legacy systems or existing enterprise applications, and then compare them with the “new” business service functions necessary in an SOA environment. This comparison between new business functions or business requirements, usually known in business process management as “to-be”, and current capabilities, known as “as-is”, gives rise to “gap analysis” (Papazoglou, 2008), a well-known concept which enables us to determine how available software service implementations may be assembled within the newly conceived business processes best suiting a business’s goals (Bolstorff and Rosenbaum, 2007), (Juan and Ou-Yang, 2004), (Nguyen et al., 2009).

Our proposal is the definition of a model-driven framework for gap analysis. The reason to adopt a model-driven approach is the benefits that Model Driven Architecture (MDA) defined by the OMG (Miller and Mukerji, 2003) adds to the development process of an SOA, and specifically to the process of recovering design information from existing software. MDA provides guidelines for capturing both the behavioural and structural aspects of an SOA using models, which could then be translated via standardized transformation rules into interface specifications for any given application platform. In recent years, however, the OMG has recognised that not all applications are developed in a top-down way, so there has to be a way of recovering design information from existing applications for which no MDA model exists (a bottom-up approach). The OMG has begun to tackle this with Architecture-Driven Modernization (ADM) (Ulrich and Newcomb, 2010).

The framework proposed for gap analysis includes: a) the conceptualization of tasks, roles and models to be analysed, and indentifying and representing as-is and to-be situations; b) the definition of a method, called GAMBUSE (Gap Analysis Method for Business Service Engineering), which includes specific steps for driving gap analysis using a model-driven approach. A previous approach of the method discussing their foundations was presented in Nguyen et al. (2009). This chapter provides also a basic implementation of the approach that illustrates a way to analyze the gap between as-is and to-be models in a particular business case.