This paper is witness to the fact that the (rigorous or quantitative) measurement and evaluation of service-oriented architectures (SOAs) is still in its infancy--even in academia. Focusing on sustainable agility, instead of the familiar reuse and cost-cutting arguments in favor of SOA, this paper clearly develops two basic research constructs and (highly superficially) evaluates the benefits provided by the SOA paradigm in five case studies.

Agility is understood in a broader sense than flexibility as the capability to adapt to unexpected changes. This is operationalized through speed/faster time to market, flexibility, proactive innovation (not evaluated), quality, and profitability.

Sustainability, the ability to meet the needs of the present without comprising the future, is assessed in the four dimensions of sufficiency (how much is enough), efficiency, consistency (between a system and its environment), and participation (stakeholder involvement).

The five case studies concisely describe drivers, enterprise architecture, and accomplishments of the respective SOA initiatives for two financial services companies, a utility, and two telephone companies (one in Germany and one in Switzerland).

The qualitative assessment of agility found that all five organizations succeeded in improving speed and time to market; flexibility improved in 80 percent, while reuse and profitability benefits could only be established in 40 percent of the cases. Sustainability was much harder to evaluate: almost 45 percent of options remained undecided, and in the other instances positive contributions to the four factors could be identified.

Nevertheless, an analysis of the results provides the most important insight of the paper: SOA as a design paradigm only contributes to short-term flexibility. In order to achieve sustainable agility, the case studies suggest that companies implement structures, processes, and instruments of enterprise architecture management (business architecture and information technology (IT) architecture). This seems even more important in light of the complexity increase brought about by SOA--another important fact established by the authors.