Dresden University of Technology
Cloud computing infrastructure services encompass temporally, spatially and structurally coarse-grained access to computing resources. This prevents a genuine pay-per-use model for the consumer and effectively leads to over-reservation and inadequate tariffs especially for short-term, small-scale computations. The authors propose an economically motivated compensation approach to increase the granularity and utility of reserved computation and storage services. The outcome is a highly-virtualising cloud resource broker. It consists of a fiduciary resource service marketplace and a consumer-configurable virtual machine for resource sharing. The system supports hierarchically nested virtualisation with dynamically adjustable resource limits for fine-grained structural, temporal and vertical-spatial scalability.