Providing proper timeliness guarantees to distributed soft real-time applications in a virtualised infrastructure involves the careful use of various techniques at different levels, ranging from real-time scheduling mechanisms at the virtual-machine hypervisor level and QoS-aware protocols at the network level, to proper design methodologies and tools for stochastic modelling and runtime provisioning of the applications. This paper describes the way these techniques were combined to provide strong quality of service guarantees to interactive soft real-time applications in the Cloud Computing infrastructure that has been developed in the con-