The cloud paradigm is currently enjoying strong interest from academia as well as businesses in view of the novel scientific challenges posed and the tremendous potential in terms of cost savings. A particularly interesting property of clouds is their
ability to scale automatically, in order to provide the service quality desired by the users for the application at hand. This scaling is possible at all layers, i.e. on the infrastructure layer, on the platform layer as well as on the application layer, and
is typically referred to as “elasticity of the cloud”. The generic support for scaling cloud components is currently limited, and the scaling logic is now located in the applications themselves (taking the decision to ask/decommission hardware and
platform components). The project proposal aims at realising an intelligent cloud middleware where the scaling intelligence is inherently provided by the middleware itself. This will not only allow faster application development (relieving application
developers from providing their own scaling logic), but will also bring opportunities to optimise across applications, yielding a better system utilisation and hence lowered operational costs.