Table 1 - uploaded by Remigijus Gustas
Content may be subject to copyright.
Source publication
The term of Enterprise Architecture (EA) has been used for many years within the information system engineering community. There are two basic challenges facing the enterprise engineering process: enterprise modelling and integration. Modelling involves system analysis of enterprise architectures that define strategic, organisational and technical...
Context in source publication
Context 1
... it is defined in various perspectives such as the "what", "who", "where", "when" and "how". The matrix of various types of enterprise architecture representations is shown in Table 1. An integrated representation of the organisational and technical system is necessary to develop a holistic understanding and to plan orderly transitional processes from the current to the target enterprise architecture. ...
Citations
... Despite of this fact, there is no complete agreement on the specific modeling principles of service architectures. Many approaches are focussing on design of services from software components by using object-oriented methods (Ambler, 2002), (Gustas & Jakobsson, 2004). Since the object-oriented models are based on computation dependent constructs, they increase complexity of a system specification. ...
This chapter presents a pragmatic-driven approach for service-oriented information system analysis and design. Its uniqueness is in exploiting a design foundation for graphical description of the semantic and pragmatic aspects of business processes that is based on the service-oriented principles. Services are viewed as dynamic subsystems. Their outputs depend not only on inputs, but on a service state as well. Intentions of business process experts are represented in terms of a set of pragmatic dependencies, which are driving the overall system engineering process. It is demonstrated how pragmatic aspects are mapped to conceptual representations, which define the semantics of business design. In contrast to the traditional system development methodologies, the main difference of the service-oriented approach is that it integrates the static and dynamic aspects into one type of diagram. Semantics of computation independent models are expressed by graphical specifications of interactions between service providers and service consumers. Semantic integrity control between static and dynamic dependencies of business processes is a one of the major benefits of service-oriented analysis and design process. It is driven by pragmatic descriptions, which are defined in terms of goals, problems and opportunities.
... EM can also be described as a generalization and an extension of system analysis and design [16]. The main idea in EM is to define a consistent, coherent and complete specification, a conceptual schema, of the future database [16] using the same schema type [17]. Applying EM during view design and view integration has several advantages if compared with a traditional modeling language. ...
View integration is a complex, error-prone and time-consuming task. Therefore there is a need to decompose the integration methods into smaller well defined phases where different techniques are applied. Most of the methods used today is composed of, or at least is a mixture of, the following four phases: pre-integration, comparison of the views, conforming the views and merging and restructuring, and most of the methods put focus on phase 2 and 3. Despite of this there is a gap between these two phases. To bridge this gap a framework has been developed. In the framework Inference Rules (IR) are used, together with Enterprise Modeling (EM) as the canonical modeling language, to deduce dependencies that are not conflicting from dependencies that are or have been identified as an inter-schema property. Three research questions regarding why, how and when IR should be applied in view integration are analyzed and discussed and it is argued that the framework may not only improve view integration as such but also improve the applied conflict and inter-schema resolution techniques.
In the field of information system (IS) design and modeling the topic of integrating different views and schemata to a common
conceptual schema is a central issue. Integration of two schemata means that they are compared, conflicts between them are
identified and resolved, and finally the schemata are merged. Integration is often based on a global ontology that provides
the valid concepts and interdependencies of a domain. In this paper we adapt the definition of ontology [8]: “An ontology is an explicit specification of a conceptualization.”. Traditional integration techniques are often based on concept name
comparison which even more motivates the use of an ontology as a domain lexicon.
In order to design databases and software components that are fulfilling the customers requirements, modeling languages that
define them at a high level of abstraction are needed e.g. [2, 12, 15, 18]. Several modeling languages and methods have been proposed but most of them put focus on the implementation level and the
technical parts of the future information system [15].
The ubiquitous use of databases (DB) in information systems (IS) today is a direct result of complex and enormous amounts
of data processing required in modern businesses. The complexity of the data processing requires software to support the organization
in being competitive in an increasingly demanding business climate. The software must support the organization in creating
business advantages, thus the need for robust, yet flexible software solutions is increasingly important to maintain or gain
effectiveness [1].
Service architecture is not necessarily bound to the technical aspects of information system development. It can be defined
by using conceptual models that are independent of any implementation technology. Unfortunately, the conventional information
system analysis and design methods cover just a part of required modelling notations for engineering of service architectures.
They do not provide effective support to maintain semantic integrity between business processes and data. Service orientation
is a paradigm that can be applied for conceptual modelling of information systems. The concept of service is rather well understood
in different domains. It can be applied equally well for conceptualization of organizational and technical information system
components. This chapter concentrates on analysis of the differences between service-oriented modelling and object-oriented
modelling. Service-oriented method is used for semantic integration of information system static and dynamic aspects.
KeywordsService-oriented modelling–Integrity of statics and dynamics–Conceptualization of events–Rules–Constraints
Enterprise models should have a capacity to describe consistently business processes across organisational and technical system boundaries. It would help system designers to understand why the technical system components are useful and how they fit into the overall organisational system. The implementation bias of many information system methodologies is a big problem for inconsistency and integrity control. The same implementation oriented foundations are often used in system analysis and design phase, without rethinking these concepts fundamentally. Common repository of most CASE tools does not guarantee the consistency of enterprise architectures for a reason that interplay among static and dynamic dependencies is not available. Enterprise modelling and integration should stick to the basic conceptualisation principle that prescribes analysis of only conceptually relevant aspects. It cannot be influenced by any implementation details. The consistency problems are best detectable and traceable at the conceptual layer. In this study on semantic dependencies, we demonstrate how various fundamental concepts from different classes of models can be interlinked and analysed together. An important result of this study is a set of inference rules. The infe- rence capability is an intrinsic feature of logical approaches, but the conventional methods of system analysis have not yet dealt in sufficient detail with the inference principles.
Information systems can be conceptualized in a number of ways. Most methodologies propose to analyze separately process and data semantics by projecting them into totally different diagram types. This system analysis and design tradition is very strong in most modeling approaches such as structured analysis as well as object-oriented design. Structural and behavioral aspects are complementary. They cannot be analyzed in isolation. Lack of a conceptual modeling approach, which can be used for verification of semantic integrity among various types of diagrams, is the cornerstone of frustration for information system architects. Inconsistency, incompleteness and ambiguity of conceptual views create difficulties in verification and validation of technical system architectures by business experts, who determine the organizational strategies. Consequently, the traditional information system methodologies are not able to bridge a communication gap among business experts and IT-system designers. Various interpretations of semantic relations in conceptual modeling approaches make the system analysis and design process more art than science. It creates difficulties to formulate comprehensible principles of decomposition and separation of concerns. Unambiguous definition of aggregation and generalization is necessary for breaking down information system functionality into coherent non-overlapping components. This article concentrates on conceptual modeling enhancements, which help to avoid semantic integrity problems in conceptualizations on various levels of abstraction. The presented conceptual modeling approach is based on a single type of diagram, which can be used for reasoning on semantic integrity between business process and data across organizational and technical system boundaries.
