Service-Oriented Architecture
ACM Classification D.2.13Traditionally, only a relatively small subset of potential computing assets has been available to developers and users. This is an inevitable result of heterogeneous, proprietary platforms and systems, a lack of federation among organizations, security concerns, and an absence of a set of standards to address those concerns.
A solution to this problem is a service-oriented model, in which highly disparate systems expose their functionality and data through a standardized set of protocols and interfaces and share them. The goal is accepting the capabilities and information of all systems, transforming it, and presenting it for consumption in a commonly understood format.
The ideal model presents a layer of services that are discrete, easily discoverable, loosely coupled, highly distributed, platform independent, synchronous or asynchronous, and secure. Once services are available in such a commonly understood and accessible format, they can be orchestrated into new types of applications and systems.
This is an enabling layer. Very few organizations have entirely interoperable subsystems, which results in inefficiency and isolation of computing assets. Applications are often inadvertently built without considering the needs of the entire organization, resulting in duplication of work already done elsewhere, possibly without incorporating important functionality developed by another group within the same organization.
A common service layer and enabling infrastructure means more rapid, intelligent development of systems with greater functionality than was previously possible. Furthermore, it allows legacy systems to coexist seamlessly with more recent deployments.
Traditionally, software applications have been considered the primary contributors of services. However, as more and more non-traditional devices such as sensors and actuators become networked, they too can contribute services.
Related Resources:<< return to topics
Contact TAP