Predictable Systems Engineering
This work package is aiming at the engineering process for software services and components. We address the challenge that to make realistic service level agreements, the actual quality of service has to be known in advance. In practice this often is unfeasible, as during design-time, neither the concrete usage profile nor access to the execution hardware is given.
Research focuses on modelling approaches which include sufficient information about the non-functional behaviour of the respective business services (i.e. Quality of Service, QoS), their implementing components and the resources they are operated on. This information is used for predicting the services’ non-functional properties. The resulting prediction services will be used in the SLA management framework (see work package WP-A5). To achieve a higher accuracy of the prediction at run-time, the prediction service and underlying model will be enhanced by run-time information. For achieving this, run-time monitoring information is required. Thus, this work package furthermore aims at providing appropriate extensions to the engineering process for realizing the required manageability for business services within an SOA.
Manageability thereby refers to a set of models, unified interfaces and infrastructural components that offer the management information and control functions required for the run-time prediction as well as the SLA management.
The following output of this work package is now publicly available:
July 2011
July 2010
May 2009
