Abstract

The topic of this thesis is exploring web services ideas from a business perspective. Much work has been done on developing the web services technology as an infrastructure for automated business interaction. Central to automated business interaction are strategic business goals and operational business processes. Some methods for creating web services systems starting at business processes are available, however they do not address the strategic business goals such a system should help to achieve. The strategic business goals web services technology may help to achieve are twofold: (1) creating new revenues by offering services through the Internet, and (2) reducing costs by improving efficiency of existing information systems. Web services systems that have to successfully operate in a business setting, should be based on an idea that is properly evaluated on its ability to achieve the strategic business goals of that organization.

An increasing number of companies and governments are interested in using web services technology to integrate their existing information systems, or provide their services via the Internet. They need a practically-oriented, light-weight approach for exploring implementation alternatives of ideas for using web services technology. Such an approach should address the various concerns of the different actors involved in such an idea. This method should start at the strategic goals to be achieved with a web services idea, and focus on evaluating whether implementation alternatives are able to reach those goals. A crucial aspect of a successful web services idea is its economic viability; every stakeholder involved should gain economic profit. Although profitability is an important strategic goal to be reached with a web services system, other strategic goals may be of concern too. Therefore, the approach should enable evaluating how requirements for achieving various strategic goals in°uence each other. At the time this research started, we found that no such approach yet existed. Also, we discovered that no such approach could be developed using just one existing requirements engineering framework. We selected two frameworks most suitable for exploring web services ideas from a business perspective of the available requirements engineering frameworks that are applicable within the topic of web services. We se- lected these two frameworks, keeping in mind the business requirements speci¯c to the web services technology. In comparing the two frameworks, we discovered that both have distinct di®erences, but share enough similarities and complement each other very well; they can be combined into one approach. In this thesis we make complementary use of the i* and e3value frameworks by combining their strong points into web services requirements engineering method called Business-oriented Approach Supporting web Service Idea Exploration (BASSIE). This approach uses i* modelling / GRL and evaluation techniques to articulate, analyze and pick alternatives for realizing web services ideas. Also, i* is used to determine how various requirements issues influence each other. The e3value framework is used to evaluate alternatives on their profitability, and it provides several viewpoints that each address specific concerns of different types of stakeholders.

BASSIE provides three viewpoints from which alternatives may be explored: (1) business value, (2) business process, and (3) information system. The business value viewpoint focuses on how value is created, distributed and consumed in a network of actors, using economic reciprocity as the main rationale. The business process viewpoint spotlights how a business idea is put into operation by showing the operational process flow and operational expenses. The information system viewpoint shows the main software components of a web services system and their interactions, as well as their operational and investment expenses. Starting from the business value viewpoint, our approach shows an iterative flow of steps for creating and evaluating models of alternatives, and using knowledge gained during the exploration of infeasible alternatives to create more viable alternatives. BASSIE prescribes, for each step in the method, how i* models are created, these models are converted into e3value models, and the e3value evaluation results are used to analyze i* models, so that model and evaluation consistency is guaranteed. For additional viewpoints, such as the business process and information system viewpoints, our approach also shows a, slightly different, iterative flow of steps. Models from these viewpoints are created using Use Case Maps (UCM), a modelling approach that provides constructs that are helpful in modelling for software engineering in a scenario-based manner. Evaluation techniques of UCM models from these two perspectives is done using e3value evaluation techniques. Therefore, these evaluations are very similar to those from business value perspective, although they only focus on economic costs and not on revenues. Finally, BASSIE combines the evaluation results from all viewpoints at scenario level into one overall profitability evaluation showing the total costs and revenues involved with an alternative. This overall evaluation is finally used to make decisions about the implementation of the web services idea.

Our hypothesis is that other frameworks may be used to evaluate a web services idea for issues not concerning economic value, in the same manner in which the e3value framework is used to evaluate the profitability of such an idea. The exploration, evaluation and selection process using other frameworks will generally be the same. The evaluation results of such a framework, together with the evaluation results from the e3value framework, are imported into the i* SR models in order to determine the impact those requirements issues have on each other.

In this thesis we illustrate our approach using an example web services idea from the domain of music rights clearance for Internet radio stations. It involved the improvement of an existing situation where web services technology provides new opportunities for revenues from Internet radio. First, the existing situation of a national rights society collecting clearance fees from conventional radio stations, and re-distributing these fees among music rights holders it represents, was evaluated in order to determine if it is sufficient for clearing rights for Internet radio stations. This resulted in several actors involved being unable to achieve their profitability goals. Following, two new alternatives were explored. The first a central clearing organization acts as a clearance proxy - resulted in a more profitable model. However, the i* evaluation showed that this model resulted in undesirable dependencies between several actors. This indicated that not only profitability, but also other, issues are might determine an alternatives viability. The second new alternative introduced a new role for the clearing organization in which it facilitates, rather than orchestrates, the music rights clearance process. From a business value perspective this alternative seemed economic viable and did not result in unwanted dependency relationships among actors. Therefore this alternative was also explored from business process and information system viewpoint, which showed that this alternative was also economically viable from operational and technical point of view. Finally, the evaluation results of all three perspectives were combined into one overall evaluation that indicated the overall viability of the alternative.

Two existing light-weight frameworks that have proofed their practical use in many various real life IT and business projects, namely i* and e3value, are the basis of our approach. For creating models from viewpoints other than the business value viewpoint, our approach makes use of Use Case Maps (UCM). All three frameworks are related to the User Requirements Notation (URN).

-- Daniel Amyot - 12 Nov 2007

Discussion

• Thesis available at http://www.cs.vu.nl/~bvdraadt/Thesis.pdf
• Please feel free to discuss this article directly on this page. Constructive comments are welcomed! Please sign your TWiki name.