| Summary: | 碩士 === 銘傳大學 === 資訊管理研究所 === 90 === In the recent years, several network topology discovery algorithms have been proposed. However, we found that these network topology discovery algorithms did not consider how to precisely bound the discovery only within the networks interesting for the network administrator. Consequently, the discovered topology information is usually only parts of networks to be managed. In common cases, the discovered topology information may contain a lot of redundant and useless networks and nodes beyond the managed network. In this thesis, we will propose several topology discovery boundary rules, which will possibly improve the previous algorithms. As a result, a more efficient and precise network topology discovery algorithm is achieved.
Considering the common topology features of most existing networks, we will propose three boundary rules as follows: (1). Domain boundary, (2). LAN/WAN boundary, and (3). IP address exclusive boundary. Based upon the above boundary rules, we will propose some more efficient network topology discovery algorithms of boundary, and we will implement the algorithm in this research.
The algorithms for topology discovery are implemented in Perl language and experimented in campus networks. The result shows that our algorithm with boundary rules has the advantage to precisely confine the discovered area. Therefore, the efficiency of topology discovery can be improved.
|
|---|
