Research On Duplicated Fault Tolerance System In The Telecom Network

Device that used in the communication network must be available in a long time. In telecom field, network-server commonly needs to adopt duplicated fault tolerance system Server System in order to ensure the high availability and affair's continuity and integrality. When the major server cannot works, the second sever can take the major server's work rapidly and ensure that calls and communication such as H.323 or H.248 will not be interrupted. Presently, the tolerance systems mainly designed for some concrete applications and belonged to application-level, cannot meet the real time demand of telecom-level network.After analyzing the main modes of the duplicated fault tolerance system currently, the author gives readers new duplicated fault tolerance system based on "layer" mode. Aimed at the short affair, high reliability and hard real-time telecom class network, this thesis implement duplicated fault tolerance system based on the most complex TCP protocol in the TCP/IP protocols, which uniformly implement the hot backup of top application based on TCP protocol. And the implementation of TCP hot backup gives an all-purpose standard mode to hot standby of other protocols. According to this mode, an all-purpose duplicated fault tolerance system based on "layer" mode from link layer to application layer can be given. This uniform application mode can save times and money to system implementing and it is beneficial to system producers or consumers.This thesis give the design and implementation of duplicated fault tolerance system based on TCP protocol and heartbeat protocol based on network layer, and explain the pivotal process and technology in detail such as synchronization of principal and subordinate. According to the test data, it shows that TCP connections be holding all along in the process of synchronization of principal and subordinate and switch of duplex computer if only both communicating terminals don't cut automatically. The duplicated fault tolerance system run good and switch rapidly in the all statuses. It indicates that the design and implementation of system is success.