Chapter8.Managing the Application Infrastructure

Chapter 8. Managing the Application Infrastructure

In the not-so-distant past, a service and an application were often the same thing. Applications were monolithic and performed a certain set of functions (services) for their users. As web-based services appeared, along with the need for extremely fast creation and modification of services for users, organizations found that assembling services from sets of interacting applications was faster than trying to build a monolithic application for each service. Modular clusters of applications can be easily assembled to support new service offerings. This reduces development costs, speeds time to market, and provides significant development leverage. Key applications, such as an order entry system, can be used by many services because it performs a common function.

Delivering superior service quality in a dynamic system composed of multiple applications and services is based on the ability to coordinate that set of supporting applications and services to meet the overall service quality goals. Understanding the relationships of all the service components and measuring the behavior of each is a major hurdle for efficient and economical service management.

This chapter discusses service quality at the highest architectural levelthat of the application infrastructure and the services provided by them that are seen directly by the end users. (Chapter 9, "Managing the Server Infrastructure," and Chapter 10, "Managing the Transport Infrastructure," discuss service quality at the web-server and the transport levels, respectively.)

This chapter's discussion of application-level service quality is in four subsections:

• The critical need to have applications designers and the network and services managers share the same perspectives about service delivery

• Application-level service metrics, which are high-level technical metrics and other end-user experience metrics

• Transaction response time, which is a primary example of the dependence of application-level service quality on lower-layer service quality

• Instrumenting web servers and other server components