Performance Management of Composite Applications in Service Oriented Architectures

Performance Management of Composite Applications in Service Oriented Architectures

Vinod K. Dubey (Booz Allen Hamilton, USA) and Daniel A. Menascé (George Mason University, USA)
DOI: 10.4018/978-1-60960-794-4.ch006
OnDemand PDF Download:
$30.00
List Price: $37.50

Abstract

The use of Service Oriented Architectures (SOA) enables the existence of a market of service providers delivering functionally equivalent services at different Quality of Service (QoS) and cost levels. The QoS of composite applications can typically be described in terms of metrics such as response time, availability, and throughput of the services that compose the application. A global utility function of the various QoS metrics is the objective function used to determine a near-optimal selection of service providers that support the composite application. This chapter describes the architecture of a QoS Broker that manages the performance of composite applications. The broker continually monitors the utility of the applications and triggers a new service selection when the utility falls below a pre-established threshold or when a service provider fails. A proof-of-concept prototype of the QoS broker demonstrates how it maintains the average utility of the composite application above the threshold in spite of service provider failures and performance degradation.
Chapter Preview
Top

Introduction

Service Oriented Architecture (SOA) enables a market of service providers delivering functionally equivalent services at different Quality of Service (QoS) and cost levels. This presents a unique opportunity for consumers to pick and choose services that meet their business and QoS needs. The selected services can be orchestrated in a process flow to optimize the execution of business processes in a cost-effective manner. We assume that service providers publish their QoS levels and offer resource reservation to guarantee them within a certain range. We also assume that the characteristics of QoS metrics, more specifically QoS levels offered by service providers, may change over time. For instance, the performance of a service may degrade due to heavy workload conditions of the associated service provider or due to some unforeseen unavailability of the service. As a result, this may affect the end-to-end QoS of a business process that may depend on such services. This requires monitoring the performance of services and business processes at runtime and, if needed, taking corrective measures to ensure that the QoS levels of running business processes were not compromised.

This chapter describes the architecture of a QoS Broker that manages the performance of composite applications. The broker facilitates near-optimal service selection for the composite application, continually monitors the utility of the applications, and triggers a new service selection when the utility falls below a pre-established threshold or when a service provider fails. A proof-of-concept prototype of the QoS broker demonstrates how it maintains the average utility of the composite application above the threshold in spite of service provider failures and performance degradation.

Complete Chapter List

Search this Book:
Reset