Abstract. Problems such as inconsistent or erroneous instrumentation often plague applications whose source code is manually instrumented during the implementation phase. Integrating performance instrumentation capabilities into the Model Driven Software Development (MDSD) process would greatly assist software engineers who do not have detailed knowledge of source code instrumentation technologies. This paper presents an approach that offers instrumentation support to software designers and developers. A collection of instrumentation patterns is defined to represent typical instrumentation scenarios for distributed applications. A UML profile derived from these patterns is then used to annotate UML models. Based on suitable code generation templates, the annotated models are transformed into instrumented source code for different instrumentation APIs. A prototypical implementation, including an adaptation to Web services, was evaluated in a lab environment.

Today, enterprise critical applications exhibit high complexity, and their non-functional properties like performance and availability are of major importance. From a user perspective, it is desirable to specify both functional and non-functional constraints on the business process model of an application system. The approach presented in this paper shows how business-level application monitoring can be performed based on implementation-level monitoring records. To achieve this, a mapping between application workflow models and an existing instrumentation is defined. This way, monitoring records, which are generated at application runtime, can be related to the workflow model. Our system uses this mapping and a stream of monitoring records to evaluate previously modelled performance constraints and to trigger reactions on constraint violations.

Abstract. Especially for the creation of Service-Oriented Architectures (SOA), Web service technologies are often the technology of choice. In this context, solutions for the management of Web services are becoming more and more important. This paper describes an approach to performance monitoring of Web services, which is based on the Application Response Measurement (ARM) standard. This approach enables generic (application source code-independent) and customizable instrumentation of synchronous, asynchronous and one-way Web service messages by attaching meta-data to messages.

Today, an increasing number of enterprise applications are based on Service Oriented Architectures (SOA). The complexity of these applications requires Service Level Management (SLM) from the business process layer down to the actual implementation of application components. However, this complexity also hinders continuous application management, which relies on monitoring non-functional properties such as performance. A major challenge is to create a unified view on the overall system including (implementation level) performance across newly developed services and legacy components. In this paper, we present a comprehensive performance-aware application design method that can be applied to newly developed applications and legacy applications. The presented method is a result of the IT management project PerManEntA 1 .

No abstract