Research On Management And Control Technology And Resource Optimization In Intelligent Optical Network

Intelligent Optical Network (ION) is considered to be the future of the next generation optical transportation network, the intelligence and dynamic in service provisioning and network management/control approach not only meets the increasing bandwidth demand of data service telecommunication service, but tackle the bursty of data traffic. At the same time, the evoluation of telecommunication service and the trend of intelligence of whole network including the service layer and edge of the network, stimulate the change of network from technology-driven to service-driven. Furthermore, it poses a challenge to intelligentization、scalability、differentiation of service provisioning, and scalability、dynamicity of network management/control technology. Within the context of this, the network management/control approach and resource optimization strategy draw a lot of attention from the industry as crucial technologies of ION.Under the support of several national grants and funding from industry area, service provision problem、integrated routing、survivable mapping and systematic fault tolerance strategy of ION are investigated in this paper, including the related work、theory extension、simulation development and so on, and some positive results are presented. The major results of this paper including:1. Regarding to the novel telecommunication service (with complex service logic and crucial QoS requirements) provisioning approach, we propose a service provision framework which depends on the interaction and coordination of management and control plane, also the implementation of policy based management and mobile agent. It could facilitate the end to end service provision with physical constraints and dynamic service management, with the purpose of resource optimization.2. We investigate the end to end resource reservation problem through revealling the challenge of previous serial signaling process in multi-domain optical network, and propose fast parallel signaling distributing mechanism——P-RSVP. The simulation results demonstrate the time efficiency of resource reservation in multi-domain LSP establishment. Besides, P-RSVP tries to avoid the resource collision between different LSP reservation approach through reducing the delay between information collection of resource avability and reservation action, and finally improve the resource avability. 3. We investigate the end to end traffic transmission delay problem in IP over WDM network, and especially for the traffic traversing more than one logical link, the queuing delay in router side plays a important rule in the traffic end to end delay. We propose an integrated routing algorithm with load balancing consideration—LBA (Load Balanced Algorithm). The simulation results show that load balancing is achieved in the whole network and the end to end delay of traffic is reduce, as well as the traffic blocking probability.4. In IP over WDM network, we arise the dynamic survivable mapping problem at first, and employ DR (Dynamic Routing) and RILT (Remapping of Instant Logical Topology) strategy for an on-line algorithm in order to remove the critical cutset in reconfigurable logical topology and improve the connectivity of logical topology under single physical link failure, so that, the affected traffic could be restored in a maximum way. The results in this paper could provide suggestion for IP layer restoration design in IP over WDM with dynamic traffic.5. In distributed computing service provisioning, IT resource and network resource should be allocated jointly. Through analyzing the systematic fault tolerance strategy requirements in practical network, we propose a heterogeneous resource control framework based on cognition theory. Also, after examining the existing problem in previous single facility node (IT resource) failure tolerace strategy, we propose the FD-EVN (Failure Dependent-Enhanced Virtual Network) based on the live migration in resource virtualization. Comparing with the previous fault tolerace strategy, FD-EVN reduces the redundancy needed to guarantee the VN to survive from single facilty node failure and improve the resource efficiency. Furthermore, the service acceptance ratio is improved and the mapping cost of VN is reduced.