PLASTIC - Test Framework: Assessment and Revision

This document is Deliverable D4.3 of PLASTIC Work Package 4 (WP4), titled: Test Framework: Assessment and Revision. It provides a detailed description of the most up-to-date version of the tools developed within WP4 along with guidelines for installing and using them. A first version of the tools with a preliminary accompanying description had been firstly released in Deliverable D4.2: Test Framework: Prototype Implementation, which can be considered superseded by this Deliverable. We refer throughout to WP4 Deliverable D4.1: Test Framework Specification and Architecture, which provided a comprehensive specification of the proposed PLASTIC validation framework, with state-of-the-art overview and justification for the adopted techniques. Although some of the basic principles underlying the framework are summoned up for completeness, the present deliverable must anyhow be considered together with D4.1. The PLASTIC validation framework is organised around two main phases, respectively called off-line and on-line. Off-line validation concerns validation at development time. In this phase services are tested in a fake/simulated environment that reproduces functional and/or extrafunctional run-time conditions. The tools made available for this stage are: 1. JAMBITION (see Chapter 2): this is a model-based testing tool that allows to automatically derive and execute invocation sequences on a service, checking whether the responses conform to a given specification, expressed as a Service State Machine (SSM). Jambition is based on a model based testing engine called Ambition , defined in D4.1, of which Jambition is the Java front-end. To facilitate the usage of Jambition, a library called MINERVA (see Chapter 3) has been embedded into the tool. Minerva permits to model SSMs via an UML modelling tool. 2. PUPPET (see Chapter 4): this is a tool to automatically generate stubs implementing external services invoked by the service under development. The mock services generated by Puppet exhibit a correct behaviour with respect to given non-functional properties. At the same time Puppet can generate stubs making invocations on the service under evaluation according to certain workload profiles. 3. WEEVIL (see Chapter 5): it consists of a synthetic-workload generator coupled with an environment for managing the deployment and execution of experiments. Weevil is intended to facilitate experimentation activities for distributed systems by providing engineers with a flexible, configurable, automated and, thus, repeatable process for evaluating their software on a networked testbed. On-line validation foresees testing of a service when it is ready for deployment and final usage. In particular, the PLASTIC validation framework supports validation during Live Usage, i.e., service behaviours are observed during real execution to reveal possible deviations from the expected behaviour. Also on-line validation can cover both functional and extra-functional properties. Tools developed to support this phase are: 1. DYNAMO-AOP (see Chapter 6): it is a framework for monitoring functional properties of external services which a BPEL process interacts with, to realize a composite service. 2. SLANGMON (see Chapter 7): it permits to dynamically detect violations of non-functional properties specified in SLAng (the PLASTIC language to specify service level specifications and service level agreements, developed within Work Package 2). Events related to the non-functional characteristics are logged and possibly used to redeem controversy.