Despite its importance, end-to-end Quality-of-Service (QoS) is still an open problem mostly because of the inter-domain segment since for the access (end-user/operator) the problem has solutions for a diversity of commercially deployed architectures. The root problem lies in the coordination complexity: operators need to coordinate the traffic services they can offer, from service choice to cooperation models. We propose a framework that allows operators to agree on a useful set of services using machine-to-machine negotiation. This consensual service set dynamically balances users' demand, effort of deployment (like cost) and local business policies - ultimately obeying economics. By having the same set of services throughout the Internet, end-to-end QoS greatly reduces to an intra-domain QoS problem. We discuss the problem using a top-down approach: from formulating a multi-dimensional, iterative multi-objective optimization problem to a design proposal based on gossip protocols.