Newly Structured Router Network Architecture Using Cloud Control Plane and Forwarding Plane

2009 
This paper proposes a novel network architecture called cloud control plane (Cloud C-plane) to reduce the power consumption of networks. In Cloud C-plane, the forwarding functions and control functions are decoupled, and the control functions are placed in the cloud. In addition, physical network topology is modified to improve the energy efficiency when the load of traffic changes. The routing engine on a dynamically reconfigurable processor (DRP) is a key component of Cloud C- plane. We implement the prototype of the routing engine on an actual DRP, and the experimental results show the execution time of the prototype is 19 times faster in the shortest path calculation. the physical topology, and the physical topology is optimized in terms of energy efficiency. Path calculation is a key in Cloud C-plane, then we em- ploy routing engine on Dynamically Reconfigurable Processor (DRP) there. This approach is based on an on-chip emulated network. The reason that we use DRP in the routing engines is to implement experiment based traffic engineering algorithm, which is a new approach to calculate a suitable path. One of the merits of DRP is flexibility. The internal circuit structure of DRP can be designed, then we make an emulated network, which corresponds to the real network, on DRP. Emulated packets are transmitted throughout the emulated network, and we can experimentally get the shortest path because the first emulated packet from the source node and the destination indicates the shortest path. We develop the prototype of the routing engine on an actual DRP. This paper as organized as follows. Section II describes the proposed architecture of Cloud C-plane. In Section III, we explain the basic algorithm of the proposed routing engine on DRP. The prototype implementation of the routing engine is shown in Section IV, and the experimental results are provided in Section V. Finally, we summarize this paper in VI.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    18
    References
    0
    Citations
    NaN
    KQI
    []