適用於FTP之檢索系統實作與分析

Abstract

隨著電腦儲存容量呈倍數快速成長，藉由磁碟陣列技術建構大容量的檔案伺服器已非難事，而如此龐大的檔案資料中，必須擁有快速的檢索機制，提供使用者全文檢索之用。小型網站或個人電腦經常採用字串逐一比對的方式來搜尋，在資料量小的狀況下，可以輕易滿足需求，但此方式必須耗費大量運算資源，已無法達到數百甚至數千萬筆以上資料幾秒內完成搜尋的效率。因此本研究中主要探討檔案伺服器上資料的檢索技術，利用現有普遍等級的電腦，研究文件結構、索引資料儲存格式與搜尋的方法，實作一個快速搜尋、精準排序的檔案搜尋引擎。

As the capacity of computer storage increases rapidly, building a large FTP server with RAID technology is neither difficult nor expensive anymore. However, we need to have an efficient retrieval method that provides a full-text search to allow the users to quickly locate what they want in such huge amount of data. Search methods that work fine on a small data set may consume too much computing resources to achieve acceptable performance when searching large databases. In this paper, we discuss data indexing techniques tailored for FTP servers and study the structure of documents, the format of indexed data and the method of searching. These efforts result in the implementation of an efficient file search engine on PC-class computers.