The Research On Key Technologies Of Signalling In Automatically Switched Optical Network

Signalling,as one of the most important function in control plane of ASON network,distinguishes the intelligent optical network from the traditional static optical network greatly.In ASON network,connections can be established, modified,configured and released in dynamically by using distributed signaling. Thus,the service can be provided and the resource of bandwidth can be allocated dynamically,which significantly reduce the configure time of new service and save the network operating costs.Many key theories and technologies are involved in the implementation of signalling in ASON network,such as distributed resource reservation schemes, signalling message crankback schemes,signalling refresh schemes,Protection and Restoration and so on.This thesis will focus on these issues and give an in-deep research.Firstly a systematic analysis is given on the characteristics of ASON distributed call and connection management architecture and signalling protocol.Then many new algorithms and mechanisms are proposed including the mechanism of distributed resource reservation,the soft state refresh mechanism based on RSVP-TE,crankback algorithms,the technology of multiple connections per a call and the mechanism of protection and restoration etc.During those researches,we developed a distributed test-bed for ASON and fulfilled relevant validation of new schemes and novel algorithms.This thesis is divided into six chapters:(1)the present status and Significance in research of the signalling technology.(2)the systematic research on ASON signalling characteristics,distributed resource reservation schemes, signalling message crankback schemes,signalling refresh schemes,and signalling message Protection etc.(3)The research on multiple connections per a call.(4)The research on protection and restoration based on distributed signalling.(5)The construction of test-bed for ASON and the experiments to verify the proposed novel algorithms and schemes.(6)Summary and expectation.The main innovations can be summarized as follows:1)Based on current crankback mechanism,a new mechanism of crankback based on re-routing in crankback node is proposed.This new mechanism is effective in reducing the frequency of crankback and reduces the final setup time for connections.Simulation results showed that,compared with current crankback mechanism,crankback based on re-routing in crankback node,the setup time of connections can be greatly reduced for more than 20%～60%;,the bandwidth can be saved about 12%～20%.Based on traditional refresh mechanism,the upper limit of local state's lifetime in RSVP-TE is defined.The local state's lifetime is not only an upper limit, there should be minimum.Its value is related with the refresh period R between generation of successive refreshes for the state by the neighbor node and the numbers of nodes in LSP.(the second chapter)2)The innovative concept that multiple connections pre a call is proposed,a call request can trigger multiple connections and these connections may complete one signaling flow;based on this concept,the novel algorithm of connections divided and routing and the new crankback mechanism are proposed.The experimental results shows that multiple connections per a call have less connection setup time and network traffic than one connection per a call have.(the third chapter)3)Failures notify mechanism based on the same source node is proposed. This mechanism classifies the connections according to same source node, and notify source node for multiple connections at one time.It will effectively reduce the times of notify,thereby increasing the network recovery time.Both theoretical and experimental simulation shows that the new mechanisms for less notify time than the traditional mechanism N/M times.4)Based on graceful connection deletion mechanism in RSVP-TE,a new mechanism of connection deletion is proposed;realize the function that the connection of PRC/SNCP deletion by called node;solve the problem that the connection of PRC/SNCP couldn't be associated by called node in distributed networks.