基於OpenStack之雲端無線電接取網路資源管理機制

Abstract

近年來，由於無線通訊產品推陳出新，手持裝置的數量將會持續上升，因此未來的無線系統將會需要更多資源。為了因應未來的趨勢，我們利用感知無線電網路 (Cognitive Radio Network) 及雲端來分配及管理巨大的資源。 在我們的論文中，我們探討了一個運作在雲端的無線接取網路(cloud-based cognitive radio access network，C2-RAN)。為了提升頻譜資源利用與提供綠能通訊，我們提出了一個基於雲、端、網的系統架構，並且在上面分別設計不同的資源管理架構。我們的資源管理架構主要分成三個部分，包括在雲端上的頻譜資源管理、在雲端上的功率控制及資源分配、以及在感知無線電接取點上的資源管理及使用者排程。在此篇論文中，我們著重於探討雲端上的資源分配，考量到C2-RAN的通訊網路服務會隨著時間的不同而使所需服務的對象有著劇烈的變化，傳統預先配置運算資源的雲端架構變的不可行，而為了解決這個問題，我們利用了雲端運算的彈性，提出一套可具動態調整雲端運算資源與平衡負載的管理方法。我們將依據使用者、AP的數目與服務範圍所產生的運算量或是虛擬機的負載，於雲端系統新增一能動態增加或減少使用計算資源的功能，並且能根據不同條件設立不同種的雲端擴展機制，藉此來提供雲端伺服器之間的負載平衡管理，以達到具節能功效的雲端系統。

In recent years, wireless communication product innovate to go beyond old ideas, and the number of mobile devices will rise daily. Thus, the future of wireless system will demand more resources. In order to respond accordingly to the trend of the future, we use Cognitive Radio Network and Cloud system to allocate and manage a lot of resources. In our paper, we discuss cloud-based cognitive radio access network (C2-RAN). In order to improve spectrum utilization and provide green communication, we present an architecture based on cloud, end-device and network, and design different resource management framework for each case. Our resource management framework is separated to three parts, clustering and resource management in Cloud, power control and channel allocation in Cloud, and resource management and user scheduling in CR access points (CR APs). In this paper, we focus on discussing the cloud resource management . Taking into consideration the radical change of the users which is led by the communication of C2-RAN vary with time, the traditional pre-configured cloud computing resources architecture becomes feasible. In order to solve this problem, we took advantage of the flexibility of cloud computing, proposed a management approach which can dynamically adjust with cloud computing resources and balance the load. We will exploit a cloud which dynamically add computing resources to the system and decrease the unnecessary virtual machine and can set up different types according to the different conditions of the cloud scaling mechanism based on the loading or the amount of the virtual machine, AP number and range of service. We provide load balancing between cloud resource management, in order to achieve a saving efficacy of cloud systems.