| Summary: | 博士 === 國立成功大學 === 資訊工程研究所 === 88 === In recent years, portable devices have been widely accepted and used in our daily life. Furthermore, wireless communication also greatly expedites the exploiting of these devices and allows networked communication mobile. On the other hand, the increasing need of wireless/mobile communication has emerged new challenges that can not be handled properly in traditional wired/fixed networks. Therefore, ATM (Asynchronous Transfer Mode) networks have been advocated as a candidate for wired/wireless transmission in the wide-area interconnection of heterogeneous networks. Today’s ATM networks have been designed to provide networking solutions for fixed network, but the use of ATM networks in a mobile environment raises many technique issues needed to be conquered yet.
This study investigated several aspects of providing integrated supports for mobile-switch ATM networks as one of the application scenarios proposed by the Wireless ATM Working Group. Our research was divided into four parts: (1) location management of mobile ATM switches, (2) topology design of mobile-switch ATM networks, (3) wireless access switch deployment and (4) resource management on an ATM switch.
In the first part, mechanisms to support the mobility of ATM switches were proposed based on the Private Network-to-Network Interface (PNNI) protocol. A limit-range topology update procedure for mobile switch tracking and a hierarchical loose-source routing algorithm for mobile switch locating were also presented. The experiment results, which showed the relationship between peer group size and network costs, triggered the following two parts of our study. In order to maintain the optimal network structure of a mobile environment in which mobile switches can move anywhere, two partition algorithms for splitting huge peer groups based on finding the congregate areas were provided. From the experiment results, we observed that our approach could lower the network overhead even in the case in which the CMR raises to 15. We introduced the Genetic Algorithm to solve the WASW planning problem in mobile-switch networks. Details of our Genetic Optimization Process were developed and two demonstrating scenarios were showed in this part. Finally, we proposed a quota-based admission control algorithm with sub-rating mechanism for resource management on mobile switches. A stochastic Petri-Net model was used to find the optimal quota values and two approximation approaches were developed to find sub-optimal settings. Based on the system workload, the approximation approaches could be used to adjust the quota values dynamically to achieve higher system performance.
|
|---|
