Skip to Main Content
This paper proposes a model to maximize revenue while serving customers with different quality-of-service (QoS) requirements. A providers' goal is to share resources between active customers to ensure all QoS requirements. At the same time, a provider is interested in maximizing the revenue. Since the amount of active users varies a provider functioning can be optimized by allocating different portions of resources. The proposed model is based on the weighted fair queue policy, which is extended so that the usage-based revenue criterion can be used to dynamically adapt weights. The model is flexible in that different services are grouped into service classes and are given different performance characteristics. It guarantees the QoS requirements and maximizes the revenue by manipulating weights of the WFQ model. The simulation of the proposed model considers a single node with several service classes. It is shown that the total revenue can be significantly improved when compared to a non-adaptive approach.
Date of Conference: 30 June-3 July 2003