Mikhail Perepletchikov

Coupling Metrics for Predicting Maintainability in Service-Oriented Designs

Service-oriented computing (SOC) is emerging as a promising paradigm for developing distributed enterprise applications. Although some initial concepts of SOC have been investigated in the research literature, and related technologies are in the process of adoption by an increasing number of enterprises, the ability to measure the structural attributes of service-oriented designs thus predicting the quality of the final software product does not currently exist. Therefore, this paper proposes a set of metrics for quantifying the structural coupling of design artefacts in service-oriented systems. The metrics, which are validated against previously established properties of coupling, are intended to predict the quality characteristic of maintainability of service-oriented software. This is expected to benefit both research and industrial communities as existing object-oriented and procedural metrics are not readily applicable to the implementation of service-oriented systems.

Cohesion Metrics for Predicting Maintainability of Service-Oriented Software

Although service-oriented computing (SOC) is a promising paradigm for developing enterprise software systems, existing research mostly assumes the existence of black box services with little attention given to the structural characteristics of the implementing software, potentially resulting in poor system maintainability. Whilst there has been some preliminary work examining coupling in a service-oriented context, there has to date been no such work on the structural property of cohesion. Consequently, this paper extends existing notions of cohesion in OO and procedural design in order to account for the unique characteristics of SOC, allowing the derivation of assumptions linking cohesion to the maintainability of service-oriented software. From these assumptions, a set of metrics are derived to quantify the degree of cohesion of service oriented design constructs. Such design level metrics are valuable because they allow the prediction of maintainability early in the SDLC.

The Impact of Service Cohesion on the Analyzability of Service-Oriented Software

Service-Oriented Computing (SOC) is intended to improve software maintainability as businesses become more agile and underlying processes and rules change more frequently. However, to date, the impact of service cohesion on the analyzability subcharacteristic of maintainability has not been rigorously studied. Consequently, this paper extends existing notions of cohesion in the Procedural and OO paradigms in order to account for the unique characteristics of SOC, thereby supporting the derivation of design-level software metrics for objectively quantifying the degree of service cohesion. The metrics are theoretically validated, and an initial empirical evaluation using a small-scale controlled study suggests that the proposed metrics could help predict analyzability early in the Software Development Life Cycle. If future industrial studies confirm these findings, the practical applicability of such metrics is to support the development of service-oriented systems that can be analyzed, and thus maintained, more easily. In addition, such metrics could help identify design problems in existing systems.

A Controlled Experiment for Evaluating the Impact of Coupling on the Maintainability of Service-Oriented Software

One of the goals of Service-Oriented Computing (SOC) is to improve software maintainability as businesses become more agile, and thus underlying processes and rules change more frequently. This paper presents a controlled experiment examining the relationship between coupling in service-oriented designs, as measured using a recently proposed suite of SOC-specific coupling metrics and software maintainability in terms of the specific subcharacteristics of analyzability, changeability, and stability. The results indicate a statistically significant causal relationship between the investigated coupling metrics and the maintainability of service-oriented software. As such, the investigated metrics can facilitate coupling related design decisions with the aim of producing more maintainable service-oriented software products.

Comparing the impact of service-oriented and object-oriented paradigms on the structural properties of software

Service-Oriented Architecture (SOA) is a promising approach for developing enterprise applications. While the concept of SOA has been described in research and industry literature, the techniques for determining optimal granularity of services and encapsulating business logic in software are unclear. This paper explores this problem using a case study developed with two contrasting approaches to building enterprise applications that utilise services, where one of the approaches employs coarse-grained services developed based on the principles of Object-Orientation (OO), and another approach is based on embedding business rules and logic into executable BPEL scripts and constructing a system as a set of fine-grained services. The quantitative comparison based on a set of mature software engineering metrics showed that a system developed using the BPEL-based approach has a potentially higher structural complexity, but at the same time lower coupling between software modules compared to an OO approach. It was also shown that some of the existing software metrics are inapplicable to SOA, hence new metrics need to be developed.

The impact of software development strategies on project and structural software attributes in SOA

Service-Oriented Architecture (SOA) is a promising approach for developing integrated enterprise applications. Although the architectural aspects of SOA have been investigated in research and industry literature, the actual process of designing and implementing services in SOA is not well understood. The goal of this paper is to identify tasks needed for successful design and implementation of services, and investigate their effect on the project and structural software attributes in the context of SOA. This facilitates the specification of guidelines for decreasing the required development effort and capital cost of the SOA projects, and improving the structural software attributes of service implementations. The tasks are identified in the context of top-down, bottom-up and meet-in-the-middle software development strategies.

A Formal Model of Service-Oriented Design Structure

Service-oriented computing (SOC) is a promising paradigm for developing enterprise software systems. The initial concepts of service-orientation have been described in the research and industry literature and software tools for assisting in the development of service-oriented (SO) applications are becoming more widely used. Nonetheless, a precise description of what constitutes a SO system is yet to be formally defined, and the design principles of SOC are not well understood. Therefore, this paper proposes a formal mathematical model covering design artefacts in service-oriented systems and their structural and behavioural properties. This model promotes a better understanding of SO concepts, and in particular, enables the definition of structural software metrics in an unambiguous, formal manner. Finally, although the proposed model is generic, it can be customised to support particular technologies as shown in this paper where the model was tailored for BPEL4WS implementation.

Prioritisation mechanisms to support incremental development of agent systems

It is often necessary to partition a project into different priority levels and to develop incrementally. This paper presents a mechanism whereby a developer can prioritise scenarios on a five point scale, leading to automated, coherent partitioning of all required design entities, according to the three IEEE defined priority levels of essential, conditional and optional, which are used in many companies. This allows for automated support to guide the developer as to what design artefacts need to be developed at each phase. The developer can indicate the relative sizes desired for the three partitions and the algorithm described will attempt to get as close to this as possible. It is also possible to move items manually to achieve better sized partitions, as long as priority orderings are not violated. The approach is fast and easy to apply at various times during development, as needed.

Exploration on software complexity metrics for business process model and notation

Business Process Model and Notation (BPMN) is a graphical representation and notation for modeling complex business processes in diagrams. A simple BPMN diagram is easier to understand by all of the business stakeholders than a complex one. It is also easier for the developers to implement the corresponding systems. Complexity metrics can measure the complexity of a diagram. Only a few BPMN complexity metrics are found in the literature as BPMN is a recent development. To propose a new BPMN complexity metric, it is important to find suitable software complexity metrics which can be further adapted to develop a complexity metric for BPMN. This research surveys the existing software complexity metrics and the existing BPMN complexity metrics (i.e. McCabe Cyclomatic Complexity, Control-flow Complexity, and Halstead-based Process Complexity Metrics) to compare their performance and suitability in measuring the complexity of BPMN diagrams. The BPMN diagrams of the business processes of two Enterprise Resource Planning (ERP) open-source systems (i.e. Compiere and Openbravo ERP systems) are used in this research. The metrics values obtained are compared with empirical application and code measurement values (i.e. number of form-fields, number of files of code, and number of classes) of the two open-source systems. This research finds that the Halstead-based Process Complexity that has been proposed in the literature is useful in measuring the data complexity of BPMN diagrams. This means that the Halstead-based Process Complexity can be further elaborated to produce a BPMN complexity measure.

Towards the definition and validation of coupling metrics for predicting maintainability in service-oriented designs

Service-oriented computing (SOC) is emerging as a promising paradigm for developing distributed enterprise applications. Although some initial concepts of SOC have been investigated in the research literature, and related technologies are in the process of adoption by an increasing number of enterprises, the ability to measure the structural attributes of service-oriented designs thus predicting the quality of the final software product does not currently exist. Therefore, this paper proposes a set of metrics for quantifying the structural coupling of design artefacts in service-oriented systems. The metrics, which are validated against previously established properties of coupling, are intended to predict the quality characteristic of maintainability of service-oriented software. This is expected to benefit both research and industrial communities as existing object-oriented and procedural metrics are not readily applicable to the implementation of service-oriented systems.