無線網路環境下依行動用戶資源特性之網際網路內容適性化策略

Abstract

目前有許多樣式的小型行動用戶手持裝置被運用於網際網路的內容存取上，如行動電話以及個人化數位助理器等。在無線網路環境中，行動用戶手持裝置在硬體及軟體的性能上呈現多樣化，如螢幕尺寸、色彩深度、緩衝器大小、運算能力、以及顯示不同資料格式的能力等。因此，如何有效地在這些不同的裝置上呈現多媒體網頁內容，使得此網頁內容能夠被行動用戶容易地瀏覽是一大挑戰。在本論文中，我們提出基於伺服器之網際網路內容轉碼系統之設計與實作。它根據我們所提出的內容適性化架構去改編多媒體網頁以便能夠呼應行動用戶手持裝置之性能。此系統有兩項關鍵元件：(1) 資料呈現結構，稱之為Info Tower：它重新組織原始的多媒體網頁內容，使成為多種解析度及多種形態之階層化表現方式；(2) 智慧型內容選取器：它用來選擇最適合之內容呈現方式，以符合裝置的性能及資源。我們提出一個內容轉碼策略以轉換文字及表格成為適合觀看之呈現方式，此乃特別是針對小型螢幕尺寸裝置而設計的。為求適應於各種行動用戶手持裝置的性能及資源，我們提出一個內容選擇策略，以便有效率地利用行動用戶手持裝置之資源。評估結果顯示，我們的方法比現有（RJC）方法有較佳的平均資源利用率。此外，我們在一個網際網路伺服器上建構我們的適性化轉碼系統，並展示如何改編多媒體內容來適應具有使用者個人設定檔的行動用戶手持裝置，如Nokia與Ericsson行動電話、PalmPilot OS 個人化數位助理器及Windows CE 平台之機器。

Many diverse and small mobile client devices, such as mobile phones and personal digital assistants, are being deployed to access Internet Web contents. Mobile client devices in wireless networks vary widely with respect to both hardware and software capabilities: screen size, color depth, buffer size, processing power, and the ability to display different data formats. It is thus a big challenge to effectively present a multimedia Web content on these different devices such that the Web content can be easily browsed by the mobile user. In this thesis, we present the design and implementation of a server-based Internet content transcoding system that follows the proposed content adaptation framework to adapt multimedia Web contents to match the capabilities of the mobile client devices that request them. This system has two key components: (1) a data representation architecture called Info Tower that reorganizes an original multimedia Web content into a multi-resolution, multi-type hierarchical representation, and (2) an intelligent Content Selector that selects the most suitable content for representation to meet device capabilities and resources. We present a content conversion strategy to transcode text and table items to have a well-viewing representation, especially for small screen size devices. To adapt to mobile client devices' different capabilities and resources, we describe a content selection policy to utilize the resource of mobile client devices efficiently. Evaluation results shows that our ERA (Efficient Resource Allocation) approach has better average resource utilization than the RJC approach. In addition, we also illustrate our adaptive transcoding framework with a Web server, which adapts multimedia contents to mobile client devices with user agent profiles using Nokia and Ericsson cellular phones, PDA with PalmPilot OS, and Windows CE machines