Product Run Metrics for service infrastructures

Service-oriented infrastructures are modern approaches to applications that lean to building distribution. These applications consist of independent software services, in which the functions and metadata can be uniformly accessed through the network. A lot of technologies may be used to implement different services that establish the basis for service-oriented infrastructures.

Business process management is an application field used in the service-oriented structure. Moreover, technologies and concepts of service-oriented infrastructure is also being widely adopted by different industries as well, thus there is a need to development quality assessment processes for service-oriented infrastructures.

A set of product –run metrics can be used to measure several qualitative aspects of service-oriented systems. These systems comprise a set of nodes of providers, with each node offering one or multiple non-mobile services. Each of these services also has one or more business functions as well. Clients communicate with these services by way of message interchange, whether asynchronous or synchronous. Asynchronous communication is usually the dominant one.

Services can also be compound or atomic. In business process management, compound services are linked to organized business processes, in which they can also be called applications to some extent. Because of the nature of compound services, they may be viewed as collaborations of single and multiple services.

Product-run metrics for service-oriented infrastructures are not mainly dependent on technology and are thus applicable for different types of service compositions and services.

In service-oriented infrastructures, it is important to analyze difference between components, objects and services in order to establish product-run metrics. Services are given out over the network and are also controlled by various parties. Communication is costly and unreliable in terns of computing resources and time. A single owner controls all components in an application that rest on a single node, which creating fast and dependable communication.

Services are also usually stateless since they indicate functions. Objects and components, on the other hand, are typically not stateless and symbolize entities. Services can be involved in a number of applications simultaneously while objects and components belong to only one application.

Moreover, services operate as processes that are separated from each other, thus partial application failure is possible. Elements that constitute component-based applications work together in a single process. This means that if one of these elements falls out, then the entire application ends. In addition, services communicate with the use of uniform protocols and data formats, while objects and components make use of propriety.

There is no service inheritance in service-oriented infrastructures since details of service implementation are outside of the scope. Thus, there are select inheritance metrics that are not applicable. However, some metrics which may be theoretically used to service-oriented software include network latency, response time and through-put. These metrics, however, become meaningless in the asynchronous communication style.

These set of metrics is not exhaustive but may be a good basis for measuring service infrastructures. Other elements of product-run metrics should also be taken into consideration as well. Another improvement method you may want to consider is the possibility of ties occurring between services.