處理器的設計朝向異質與多核心發展，每一個核心更可以擁有獨立的電壓與時脈控制機制，如何調整虛擬化平台的排程器(Scheduler)，來進行有效的資源管理以達到最佳的運算效能與功耗表現，是新一代的虛擬化平台設計的顯學。虛擬化平台主要為一個虛擬機監控器，透過其對電腦系統的虛擬化，提供同時可執行多個虛擬機器的目的。隨著雲端運算技術的發展，世界各地先後都建立起資料中心，使能源消耗也日益遽增，在電費不斷高漲的時代，如何節能已成為相當重要的議題。要達到資料中心省電，工作排程是一個有效的方法；另外，於行動裝置上，因智慧型手機已經成為人人必備的隨身裝備之一，促使許多不同的開發商紛紛投入手機系統與應用程式的開發，但越來越豐富的功能也使得手機能源消耗增加。本文提出以 Microkernel 為主之系統虛擬化技術，針對工作耗電情況，設計一可採用處理器具備的動態電壓與頻率調整技術之高效率排程器， 此技術不但可應用於資料中心之工作合併排程，以達到節能的目的。在行動平台上，亦可以解決不同手機系統無法相容的問題，同時提升能源使用效率並延長行動裝置使用時間。本文將介紹如何在Microkernel 上實現系統虛擬化以及支援動態電壓頻率調整之省電排程器的設計及實作方法。本文中所提方法的特色為將排程器與動態電壓頻率調整機制整合，可比現有排程與調頻分開設計的做法節省更多能源。

Heterogeneous multi-core processors have become the mainstream in modern processor design. In such platforms, each core may even have the capacity to tune its voltage and frequency independently. How to design an energy-efficient scheduler for virtualized heterogeneous multi-core platforms has become a critical issue both in the academics and in the industry. Virtual Machine Monitor (VMM) is the key component to achieve platform virtualization. VMM virtualizes several computer components to allow more than one virtual machines to execute on one physical machine simultaneously. As the cloud computing technologies keep enhancing, there are more and more new data centers in the world which results in more energy consumption.How to save energy becomes an important issue. To achieve power saving in data center, energy-aware resource scheduling is an effective solution. For mobile devices, smart phone have become an important part of our daily life. Most smart phones run only one operating system. However, there are occasions that require different operating systems on one Smartphone. Virtualization is the technology for sati sfying these requirements. Since new CPUs have the ability to turn off or adjust the core voltage and frequency of individual cores, each core can have different computing powers and energy consumption according to needs.
In this paper, we present an energy-efficient microkernel solution for virtualization. We also propose an energy-efficient scheduling module which leverages per-core dynamic voltage and frequency scaling to reduce
energy consumption. Unlike most existing techniques that handle task scheduling and voltage/frequency scaling separately, our scheduler takes both factors into consideration simultaneously. As our experiment
result shows, our approach reduces more energy consumption without sacrificing application requirements,when compared with the state-of-the-practice scheduling method.