A multi-layered traffic engineering architecture for the electronic/optical hybrid network

2003 
The electronic/optical hybrid network deploys the integration of optical GMPLS technology and IPv6 routing technology [J. Murayama, Aug. 2003]. This hybrid network improves the scalability from the viewpoint of the number of PE (provider edge) routers. Additionally, a high performance of packet forwarding is also respected. To improve forwarding performance of this network, a cut-through optical path that directly connects a pair of PE routers with high traffic demand is provided. This path improves a performance of packet forwarding between the PE routers, and avoids traffic congestion at electronic P (provider) routers. Since a number of cut-through optical paths are limited according to a limitation of network resources, a cut-through optical path should be assigned efficiently. This paper proposes an optimum path allocation algorithm according to traffic demand and resource limitation. On the other hand, to keep a best forwarding performance even if traffic demand changes, the reassignment of cut-through optical paths is needed. Such a reassignment could degrade quality of packet forwarding because a traffic concentration at an electronic P router is caused at a substituting time. This paper proposes a traffic control mechanism to avoid a traffic concentration at an electronic P router. To control a constant traffic concentration, we propose a load balancing method between electronic P routers. To control a temporal traffic concentration, we propose a cooperative operation between a cut-through optical path reassignment and IPv6 path assignment. This paper proposes a multi-layered traffic engineering server (MLTES) that provides the proposed functions and manages optical paths and IPv6 paths cooperatively and efficiently. MLTES achieves a high performance of an electronic/optical hybrid network, and provides a low cost and high performance IP-VPN platform.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    3
    References
    1
    Citations
    NaN
    KQI
    []