The federation of authority in global public computing systems poses major resource management challenges, as different stakeholders may have different views on how server resources are to be apportioned. Ultimately, the complexity of declaring and managing potentially overlapping federated policies often leads to the absence of high-level resource management systems from global public computing platforms. In this paper we propose a practical system that allows the different stakeholders to independently express federated policies, provides mechanisms for resolving potential constraint overlaps automatically, and reaches decentralised resource allocation decisions. We demonstrate experimentally that the system scales gracefully, introduces only a very low performance overhead, and is suitable for operating in realistically large and complex settings.