An engineering approach to QoS provisioning over the Internet

Existing communications systems are rapidly converging into an ubiquitous information infrastructure that does not distinguish between computing and communications, but rather provides a set of distributed services to the user. The research community must be prepared to foresee these changes and to deal with them, enlarging the space of technical possibilities so as to make available to society's needs new valuable choices. In this scenario the capability of the network to provide the applications with end-to-end Quality of Service (QoS) becomes a central issue. An engineering approach is needed in this research field in order to incrementally build the next-generation network. This paper focuses on some of the hot topics related to end-to-end QoS provisioning over the Internet and aims at exploiting the current proposals of the research community, while looking at them from a critical point of view and providing actual implementation of some of the discussed ideas. Thus, we propose a QoS-capable architecture aiming at providing flexible and effective implementation of the Integrated Services model via a Weighted Fair Queueing scheduling mechanism, while defining a new service class capable of giving long-term rate guarantees to Internet flows.