智慧型網域名稱系統規劃與管理系統之研製

Abstract

網域名稱系統是網路基礎建設中不可忽視的一個環節。在建立網路實體連線之前，常常需要一些網域名稱的查詢與轉換。然而這個重要的系統卻因為它階層式與分散式的架構以及領域專家的缺乏，使得系統管理者在管理與規劃網域名稱系統時遭遇到很多的困難。 在此篇論文中，我們提出一個智慧型的網域名稱伺服器規劃與管理系統(iDNS-PMS)來幫助系統管理者處理所遭遇到的規劃與管理上的問題並解決領域專家不足的困境。藉由一個系統化的知識擷取演算法(DNS-OCA)，領域專家的知識可被萃取出來並建構成一個網域名稱領域的本體結構，這樣的本體結構除了可以用來幫助了解網域名稱領域的知識外，還可以用來建構知識庫中的基本單位－網域名稱系統知識物件(DNS-KO)。經過知識擷取的程序，我們建立了一個經過專家驗證的網域名稱領域本體結構。最後我們實作出一個iDNS-PMS的原型系統來評估實作上的可行性，並作為後續研究的參考。

The Domain Name System (DNS) is an essential part of the Internet infrastructure. Today, most internetworking services are based on the working model that there will be some DNS queries before the communication activities. The shortage of DNS domain experts, and the hierarchical and distributed properties of the DNS system make the planning and management more difficult. In this thesis, we propose an Intelligent DNS Planning and Management System (iDNS-PMS) to help inexperienced administrators solve the planning and management problems. DNS Ontology Construction Algorithm (DNS-OCA) is a knowledge acquisition algorithm that can be used to systematically extract knowledge from domain experts to construct the DNS ontology hierarchy. The DNS ontology hierarchy can help administrators understand the DNS domain knowledge and can be used to construct the foundation of the knowledge base, DNS Knowledge Objects (DNS-KO). We construct the DNS ontology after we perform the DNS-OCA, and it has been verified by the experts. It is not only useful for people to understand the DNS domain but also helpful when we construct our system. Finally, We implement a prototype of iDNS-PMS to evaluate the potential of this system and for advanced research in the future.