資源有限嵌入式設備上之可延展式系統之研究

Abstract

隨著嵌入式系統與網際網路技術的快速發展，具有網路功能的嵌入式設備已愈見流行。這些設備在設計上有兩大趨勢。一、嵌入式設備已漸漸由靜態且只具特定功能的系統轉變為動態且具擴充性能的平台。二、因為Java的優越特性，嵌入式系統研究人員開始致力於將Java移植到嵌入式的平台上。然而，由於嵌入式設備的資源有限，不論是讓嵌入式設備具有可擴充性或是將Java移植到嵌入式平台上，都將是一項挑戰。 在本論文中我們詳述了為了有效率地達成上述兩項目標而做的努力。首先，我們提出了作業系統入口架構。它讓一般嵌入式的核心程式變得具有可擴充性。同時，它也嘗試將所增加的額外負擔保持在最小。在此架構中，我們將核心程式模組儲存在資源豐富的伺服器上(即作業系統入口)，並且在需要用這些模組時才將它們讀取到嵌入式設備中。如此一來，我們就不需要在這些嵌入式系統上加裝額外的儲存設備了。除此之外，我們也提出了方法來降低嵌入式系統上記憶體的使用量。並且為線上核心程式模組置換提供了一個解決方式。 再者，我們提出了經濟型的Java虛擬機器。它是在具有網路功能且資源有限的嵌入式設備上的一個Java程式執行平台。我們利用分散式Java虛擬機器的架構，將所有的Java程式碼儲存在伺服器上以降低嵌入式設備對於儲存容量的需求。此外，我們提出了兩項技術來降低其記憶體需求。這兩項技術分別為：伺服器端類別表示法的轉換，及隨選Java程式碼的讀取。最後，我們在嵌入式設備上用一部份的記憶體來作為Java程式碼的快取記憶體，並且分析不同的快取策略對系統效能的影響。 根據實驗量測結果，我們所使用的技術顯著地降低了嵌入式系統的資源需求。這提昇了我們架構的可應用性，使我們的技術可以應用於各式各樣的嵌入式設備上。

With the rapid development on embedded system techniques and Internet technologies, network-enabled embedded devices has grown in their popularity. Two critical design trends of such devices are as follows. First, embedded devices are shifting from static, fixed-function systems to more dynamic and extensible ones. Second, owing to the excellent features of Java, embedded system researchers start seeking ways to make these devices Java-enabled. However, making embedded devices extensible and applying Java technology to these devices are both challenging due to the shortage of resources on these devices. In this thesis, we present our efforts on achieving the above two goals. First of all, we describe the OSP framework (Operating System Portal framework), which makes embedded kernels become extensible while keeping the added overheads minimal. By storing kernel modules on a resource-rich server (i.e. the OS Portal) and loading them on demand, the need for equipping a local storage on the device is eliminated. In addition, we propose mechanisms for reducing the memory requirements and performing on-line module-replacement on the embedded devices. Secondly, we present EJVM (Economic Java Virtual Machine), an economic way to run Java programs on network-enabled, and resource-limited embedded devices. Espousing the architecture proposed by distributed JVM, we store all Java codes on the server to reduce the storage needs of the client devices. In addition, we use two novel techniques to reduce the client-side memory footprints: server-side class representation conversion and on-demand bytecode loading. And we maintain client-side caches and provide performance evaluation on different caching policies. According to the performance measurement, our techniques reduce noticeable resource requirements of the embedded devices. This allows our framework to be applied on a wide range of embedded devices.