Utility based service differentiation in CDMA data networks

  • Authors:

  • Haitao Lin;Mainak Chatterjee;Sajal K. Das

  • Affiliations:

  • Wireless Core Systems Engineering, Nortel, Richardson, Texas;Department of Electrical and Computer Engineering, University of Central Florida, Orlando, FL;Center for Research in Wireless Mobility and Networking (CReWMaN), Department of Computer Science and Engineering, The University of Texas at Arlington, Arlington, TX

  • Venue:
  • Wireless Networks
  • Year:
  • 2006

Quantified Score

Hi-index 0.00

Visualization

Abstract

The wireless data services are getting more and more competitive because of the presence of multiple service providers, all of whom offer some relative advantages and flexibilities over the others. As a result, the user churn behavior (i.e., migration from one service provider to another) is causing tremendous revenue loss for the service providers and also failure of existing resource management algorithms to fully capture the impact of churning. Moreover, the quality of service (QoS) offered to users belonging to different classes calls for new resource management schemes that address the issues related to differentiated services. In this paper, we propose a framework to study the impact of user churn behavior on the resource management algorithms to provide class-based differentiated services in CDMA data networks. In particular, our framework incorporates the user churning behavior into the admission control and power management algorithms, so that the service provider's revenue loss due to churn can be minimized. Since optimal rate/power allocation in multi-rate CDMA systems is in general NP-Complete, we provide heuristics that try to provide solutions to the resource allocation problem in real-time. In our proposed framework, we add another layer of power management called Class-Based Power Allocation/Reduction (CBPAR) function, which works with the rate control algorithm to achieve power allocation. With CBPAR, the number of variables of the optimization problem is significantly reduced helping achieve the results in real-time. Our simulation study shows that the service provider's revenue can be improved with the help of CBPAR framework. It also reveals the relationship between users' sensitivity and tolerance to QoS degradation and optimal power allocations.