Toward an on demand service-oriented architecture

IBM Systems Journal,  March, 2005  by C.H. Crawford,  G.P. Bate,  L. Cherbakov,  K. Holley,  C. Tsocanos

• E-mail
• Print
• Link

A second set of policies defines guiding principles that are focused on management of the environment with respect to how the organization, processes, and tools are defined, developed, and built. This is required in order to sustain a well-operated environment. Separating the processing and data environments from the service-development environment introduces a series of issues that must be addressed. The separation of business, application, and IT policies also requires that various issues be correlated, rationalized, and arbitrated.

DESIGNING AN ON DEMAND SOA

There are multiple techniques that can be used to perform business transformation. Transformation implies that the re-engineering of process workflows does not end solely with a successful implementation of a system or education program. Rather, business transformation has many broader elements, including those handled by organization and change management competencies.

Previously, business strategy and analysis techniques focused on a process-centric model of a business. The process view did not always force the generalization of common tasks, nor did it aid in identification of "specialist service centers"--the building blocks of on demand computing. The emerging IBM-developed component business modeling (CBM) technique represents a business as a set of collaborating components that consume and provide business services. These business components are combined to form a viable operating model. Because SOA has implications for both the technology and the business levels, the transformation of CBM-identified business services into an on demand SOA brings more business focus to how we discover and publish SOA services.

On demand SOA requires the key elements of organizational transformation to be in place. Otherwise, the culture of the organization can continue to be a major obstacle in adopting new processes, technologies, or management techniques. In addition, an on demand transformation is much broader than any tactical initiative and needs full management leadership and commitment from the highest levels down. Dynamic process and technology sourcing decisions will not make employees comfortable without appropriate communication and education. As a result, business transformation projects are often put at risk, with employees potentially resisting any change and in some cases even acting as saboteurs to the adoption of new processes. In addition, technological constraints, whether purposefully implemented or due to current technology limitations, can impose barriers to meeting certain performance objectives. (Performance objectives are typically metrics incorporated in business processes in order to measure key attributes such as speed, cost, and accuracy of results.)

Business process integration is a business enabler which requires that an on demand SOA be in place in order to provide the necessary workflow infrastructure. Today, application development tools have evolved to a point where process life cycles can be enabled from start to finish. Moreover, process-modeling tools can automatically generate use cases and interaction models. These models can be dynamically instantiated with components provided as part of the application framework. Applications are then assembled from a collection of services that provide specific functionality. Ideally, this functionality becomes part of an extensive library that can be published as Web services, which in turn can be reused by others simply by discovering the service by means of the appropriate service directory. The benefit of this model is that it is based on the application of frameworks, which are being improved on a daily basis to add increased breadth of packaged functionality. Admittedly, however, this somewhat utopian view of services still lies in the future on the business transformation roadmap. Indeed, the discussion clearly shows the large gap between what we have today and what is required in the future.