Pricing the quality of differentiated services for media oriented real-time applications: a multi-attribute negotiation approach

Abstract

The paper proposes a new algorithm for negotiating the price of quality of service in IP differentiated services networks. We focus on real-time media-oriented applications such as video conferencing, online gaming, broadcast TV and live events, streaming video, and audio on demand. The performance objectives of end-users are expressed as the end-to-end delay thresholds to be exceeded with a given maximum probability. The target of negotiation is a multi-attribute description of traffic profile and quality of service, rather than the simple raw-bandwidth attribute. The service class chosen for each traffic flow is the result of negotiation and depends on the user's and supplier's utility and quality functions and on their conceding versus selfish negotiation attitude. We model non-linear utility and quality functions in such a way to represent the user's and supplier's perception of quality of service parameters. This represents a fundamental contribution of this paper with respect to current approaches accounting for simple linear utility functions of the raw-bandwidth attribute. We analyze the utilization of network resources as well as the customer's and supplier's utility through simulation by comparing our algorithm with previous algorithms negotiating raw-bandwidth instead of end-to-end quality of service.