Abstract

The evolution of the Internet as an important communications infrastructure for the public and the convergence of various network services and applications over the Internet have been intensifying the demand for QoS-support in the Internet. The main goal of QoS-support in the Internet is to provide assured network performance relative to the various applications’ needs at reasonable incremental cost as compared with a best-effort network in a way that is scalable.
In this paper, we will address the economic problems of network resource management over more generalized hybrid-QoS networks such as IntServ-DiffServ; IntServ-IP/ATM; DiffServ-IP/ATM; Overengineering-DiffServ. All of these assume that DiffServ and IP/ATM are scalable enough for backbone network transport. It is unlikely that all the backbone network providers will employ the same QoS-mechanisms. In the expected scenarios of these hybrid-QoS networks, one of the major QoS-management problems will be how to allocate the QoS requirement and the network resources among the various carriers (ie, how much QoS is “outsourced” by a carrier to another). One way of such outsourcing of QoS is to change the service mapping between service providers. This paper will demonstrate major tradeoffs of managing the end-to-end QoS of the various hybrid-QoS networks interconnected with the Internet.

Date

2001

Authors

Junseok Hwang, Martin BH Weiss, LW McKnight, J Wroclawski

Publisher

Internet Services: The Economics of Quality of Service in Networked Markets, MIT Press