Virtual Traffic Path Optimization in Connection-Oriented Networks with Stochastic Traffic

  • Authors:
  • Dongfang Zheng;Xian Liu;Mrinal Mandal;Weidong Lu

  • Affiliations:
  • Multimedia Computing and Communications Laboratory, Department of Electrical and Computer Engineering, 2nd floor, ECERF, University of Alberta, Edmonton, Canada T6G 2V4;Department of Systems Engineering, University of Arkansas at Little Rock, Little Rock, Arkansas, USA. E-mail: xxliu@ualr.edu;Multimedia Computing and Communications Laboratory, Department of Electrical and Computer Engineering, 2nd floor, ECERF, University of Alberta, Edmonton, Canada T6G 2V4. E-mail: mandal@ece ...;Multimedia Computing and Communications Laboratory, Department of Electrical and Computer Engineering, 2nd floor, ECERF, University of Alberta, Edmonton, Canada T6G 2V4

  • Venue:
  • Journal of Network and Systems Management
  • Year:
  • 2004

Quantified Score

Hi-index 0.00

Visualization

Abstract

Traffic control is a critical issue in connection-oriented packet-switching networks such as asynchronus transfer mode, Mutiprotocol Label Switching, and Internet Protocol with IntServ. In this paper, we present a generalized concept, the virtual traffic path (VTP), to characterize the traffic control problems in connection-oriented networks. The VTP distribution typically addresses logical network design based on the physical network, and involves both the call level and the flow level controls. To date, various VTP optimization schemes for connection-oriented networks have been proposed. However, most reported schemes are based on the conventional flow assignment model. In this paper, we propose an extended flow assignment model focusing on the connection-oriented service with a non-linear objective function. The proposed model incorporates two concepts: VTP capacity and VTP flow, to perform the optimization. This model distributes traffic on all available VTPs evenly and takes the redundant capacities into account. In addition, we introduce a stochastic programming methodology to allocate VTPs when the injected traffic changes stochastically. Experimental results show that the proposed model and the stochastic methodology can significantly improve the performance of networks.