Internet applications and users have diverse quality of service expectations, making the existing same-service-to-all model inadequate and limiting. There is a widespread consensus that the Internet architecture has to be extended with service differentiation mechanisms, so that certain users and applications get a better service than others at a higher cost.
One approach, referred to as absolute differentiated services, is based on admission control and resource reservations in order to provide guarantees for absolute performance measures, such as a minimum service rate or a maximum end-to-end delay. Another approach, which is simpler in terms of implementation, deployment, and network manageability, is to offer relative differentiated services between a small number of classes of service. These classes are ordered based on their packet forwarding quality, in terms of per-hop metrics for the queueing delays and loss rates, giving the assurance that higher classes are better than lower classes. The applications and users, in this setting, can dynamically select the class that best meets their quality and pricing constraints, without a priori guarantees for the actual performance level of each class.
In the context of relative services, the Proportional Differentiation Model aims to provide the network operator with the `tuning knobs’ for adjusting the quality spacing between classes, independent of the class loads, and even in short timescales. When this quality spacing is feasible, it can lead to predictable (from the user’s perspective) and controllable (from the provider’s perspective) class differentiation, which are two important features for relative services. The proportional differentiation model can be implemented with simple and scalable router mechanisms for packet scheduling and buffer management.
In this talk, I will summarize our work on the proportional differentiation model, on the related router mechanisms for scheduling and buffer management, and on the provisioning of proportional differentiated services.