9th International Conference on Communications and Networking in China

Research Article

On the Knapsack-Based Resource Allocation for Reconfigurable Network Architecture

  • @INPROCEEDINGS{10.4108/icst.chinacom.2014.256283,
        author={Kai Pan and Weiyang Liu and Hui Li},
        title={On the Knapsack-Based Resource Allocation for Reconfigurable Network Architecture},
        proceedings={9th International Conference on Communications and Networking in China},
        publisher={IEEE},
        proceedings_a={CHINACOM},
        year={2015},
        month={1},
        keywords={resource allocation; knapsack problem; qos; reconfigurable network},
        doi={10.4108/icst.chinacom.2014.256283}
    }
    
  • Kai Pan
    Weiyang Liu
    Hui Li
    Year: 2015
    On the Knapsack-Based Resource Allocation for Reconfigurable Network Architecture
    CHINACOM
    IEEE
    DOI: 10.4108/icst.chinacom.2014.256283
Kai Pan1,*, Weiyang Liu1, Hui Li1
  • 1: Peking University Shenzhen Graduate School
*Contact email: pankai@pku.edu.cn

Abstract

Though the decades of great success the Internet has experienced, it still confronts with many challenges to achieve QoS of various business types. A consensus has been reached that resource allocation presents a tremendously status in such a network bandwidth poor settings. In this paper, we bring the knapsack problem idea into reconfigurable network architecture (RNA), which is proposed with the intention of achieving QoS of various business types, to study the resource allocation problem. By defining the volume of flows, we argue that a flow with smaller volume should be granted higher priority while it is allocated bandwidth. Compared to the conventional approach, our method is good at maximizing network utility and decreasing overall price under certain conditions. Simulation results show that our method retains the soundness of the conventional approach in the scenario of increasing flows network, and outperforms it both in the aspect of overall price and load balance under network with variation of flows.