With the development of Internet technology, the network services become more complex and various. Multicast network has been widely applied for the data transmission. It will be the one of the main direction of network research. Since the complex situation of current internet, the traditional Multicast has many problems in reliability and scalability. Multicast based on hypercube has been given more and more attention for its reliability, fault tolerance, maintainability and scalability of the multicast protocols. By the way, as the rapidly increment of the node number in Internet and the randomness caused by nodes' add-in and deleted from the network, reliable multicast service are urgently needed to ensure the reliability of data transmission. And the research of multicast based on hypercube became the hot point of current Internet technology.This thesis analyzes the advantage and disadvantage of current hypercube multicast protocol based on the basic conceptions and some related works in domestic and overseas areas. Hypercast-plus based on Hypercast and CubeFullDist is presented to solve the problems of multicast reliability. Hypercast-plus gives the complete mechanism of error control in the research of reliable hypercube multicast transmission.Due to the better fault tolerance ability of hypercube multicast and the control topology requirements of the reliable multicast transmission, we provide control model of reliable multicast based on logical hypercube, and we also provide protocol of the reliable and fault tolerance multicast based on the analysis of the performances of the model. Since nodes in multicast may leave or join at any moment, it is difficult to maintain the information. The mechanism is implemented by saving the information of the node's neighborhoods and adding logic links when the physics links are missing. According to the theory analysis and simulations, Hypercast-plus protocol shows more efficiency and reliability.Each node in hypercast-plus is defined as a local recovery unit. And some adjacent local recovery units constitute a local feedback group. Forward Error Correction (FEC) is used to recover the missing data. When the multicast message is not received, nodes will send Negative Acknowledgment (NAK) to the local feedback group for data recover. After the local feedback group receives the corresponding request ,it will send message again. If the local feedback group can not recover, it will send Automatic Repeat Request (ARQ) to the source node. Both the theory analysis and simulations show that error control mechanism can satisfy the requirement of the data transmission reliability, when the packet loss rate was under 10%, the average of packet received rate can reached above 95%. |