Ethernet Physical Topology Auto-discovery Algorithm To Achieve

The increasing number of nodes in Ethernet networks leads to the enlargement on the scale of the networks. Due to the dynamic nature of modern IP networks, precise discovery of the physical topology cannot be accomplished by tracking the elements of the networks manually. For the purpose of daily network management tasks, it is essential to acquire accurate and up to date network topology information. As a result, there is a great demand for the mechanisms and algorithms that are capable of automatically discovering the physical topology of Ethernet networks and are viable for commercial applications for the current market.In this dissertation, first of all, the necessities of automatic discovery of the physical topology of Ethernet networks are introduced. Secondly, relevant network management knowledge is overviewed and existing techniques for topology discovery are introduced and analysed comparatively. Finally, the proposed automatic topology discovery algorithm is described in detail.The current work related to the development of automatic topology discovery algorithms is mainly based on MIB or rely on end-to-end measurements. These techniques can provide the logical topology that is used for forwarding but cannot discover inactive links. Hence, the algorithm using RSTP get message to realise the discovery of both active and inactive links and eventually generates accurate topology information, is a promising solution for commercial applications.The algorithm has been tested in a lab network with switches from multiple vendors. Preliminary results verify the accuracy and robustness of the algorithm, which also indicate its potential for commercial implementations.