A Study of Genetic Based Algorithm for Managed IP Network Management

• Main Authors: Hsu-Wei Li, 李序韡
• Other Authors: Chien-Pen Chuang
• Format: Others
• Language: zh-TW
• Published: 2003
• Online Access: http://ndltd.ncl.edu.tw/handle/23574491966467189913

碩士 === 國立臺灣師範大學 === 工業教育研究所 === 91 === MPLS (Multi-Protocol Label Switching) is a new Connection Oriented Network communication technique type. It can not only transfer the data format between different network communication hardware and make them be used in the same network system but also have controlltion and management ability for network. MPLS can reduce the burden of manager and improve the transmission and transformation speed in network systems. But OSPF (Open shortest path first) routing algorithms using by MPLS is lack of consideration for the real network. This makes the path decision get into a jam easily. So the research will focus on MPLS routing algorithm, and recommend that using Genetic Algorithm to instead of traditional OSPF algorithm. Listing a routing table of different demands like normal routing, dual-homing routing, and bandwidth reservation routing. After collating and analy the experimental result we have the conclusions as bellow. (1). Using Genetic Algorithms to instead of OSPF in MPLS routing and traffic engineering can support QoS mechanism in routing algorithm and improve the ability of the transmitting path selection in MPLS network. This part can provide for the future network development and professionals for reference. (2). For the routing. If the number of data stream is lesser, the transmitting efficiency is better at dual-homing demand routing in network; if the percentage of bandwidth reservation is become higher, the transmitting efficiency will become worse.