Prediction of non-functional properties of IT-services

Prediction of non-functional properties of IT-services
Project members: Dr.-Ing. Sascha Bosse
Project status: running

The service-orientation of the IT-branch is of increasing importance, not least because of the concept of cloud computing. It supports the commoditization of IT and increases the transparency for the clients. For the providers of such services, the challenge is to offer IT-services with minimized costs and sufficient quality. Service-Level-Agreements (SLAs) are important secondary conditions. In addition to functional and legal aspects, these agreements also give non-functional guarantees of quality, for example to availability or response time of a certain service. Violating those can lead to contractual penalties or even to a loss of reputation in respect of the clients and thus, should be avoided.

In the context of the development of new and improved services and underlying IT-system landscapes, it should be aspired to predict the non-functional features of these services in advance without incurring significant costs. Therewith, the negotiation and supervision of SLAs could be supported. Furthermore, the service costs could be analyzed on different quality levels.

In order to offer such a decision support, a framework for the prediction of non-functional features of IT-services is developed in the context of this research project. This framework is based on a discrete event simulation and considers dependencies among the single properties such as the dependency among security standards and response time or of availability and throughput.