服務功能鏈問題的移動感知演算法

Abstract

網絡功能虛擬化（NFV）現在已經成為一個熱門的話題，其主要是將傳統上高度整合在硬體設備中的服務，透過虛擬化的技術應用到商用服務器。受益於這一概念，服務功能鏈（SFC）已然成為當今新應用程序構建方式的主流，將巨大且資源消耗大的網絡服務分解為一系列微小的，鬆散連接的虛擬網絡功能（VNF） 。通過這種方式，服務提供商可以更靈活地配置自己的服務。在此背景下，移動邊緣計算（MEC）的概念被提出討論。 MEC可以在無線接入網絡（RAN）內為移動用戶提供IT服務環境和雲端計算功能。目的是減少延遲，確保高效的網路運營和服務傳遞品質，並提供改進的用戶體驗。然而，隨著移動用戶的爆炸性增長，越來越多的對於延遲敏感的服務，如車載網絡，增強實境......等，給傳統的雲端環境帶來了壓力。這些服務通常伴隨著頻繁的用戶移動和對延遲的嚴格限制，這意味著雲端服務不能離用戶太遠。為了妥善處理這種困境，需要MEC遷移機制來根據不同的服務需求來管理用戶服務。服務實時遷移允許將連續運行的VNF從一個MEC移動到另一個MEC，這使得可以確保這些用戶的體驗品質（QoE）。儘管服務實時遷移技術使其成為可能，但它卻大幅的增加頻寬上的使用。在本篇論文中，我們提出了一種有效的演算法，以充分利用服務器的資源來容納更多用戶，同時保證系統中每個用戶的無縫服務體驗。模擬結果顯示，與其他算法相比，我們的演算法可以在相同的服務接受率下最小化服務遷移時間。特別是，我們的模擬顯示我們的演算法可以在滿足未來服務遷移請求的同時獲得更好的資源利用率。

Network Function Virtualization (NFV) has now become a promising concept which virtualizes hardware based application to commodity servers. Based on this concept, Service Function Chaining (SFC) is becoming the de facto approach to how new applications are built today, which a giant and resource-hungry network service is decomposed into a chain of tiny, loosely connected virtual network functions (VNF). In this way, service provider can deploy their own services with more flexibility. In this context, the concept of Mobile Edge Computing (MEC) has been proposed. MEC provides an IT service environment and cloud-computing capabilities at the edge of the mobile network, within the Radio Access Network (RAN) and in close proximity to mobile users. The aim is to reduce latency, ensure highly efficient network operation and service delivery, and offer an improved user experience. However, with the explosive growth in mobile users, more and more delay-sensitive services, such as vehicle network, augmented reality… and so on, put stress on traditional cloud environment. These services usually come with frequent user movements and harsh restriction on latency, which means the service user requests can not be placed too far from the user. To handle this dilemma properly, there is a need for MEC migration mechanism to manage user services according to the different service requirement. Service live migration allows moving a continuously running VNF from one MEC to another, which makes it possible to ensure Quality of Experience (QoE) for these users. Although the technique of service live migration makes it possible, it put a great stress on bandwidth usage. In this work, we propose an efficient online Follow-Me Chain (FMC) algorithm to fully utilize the server's resource to accommodate more users while guaranteeing seamless service experience for every user in the system. The simulation results show that FMC can minimize service migration time with same service acceptance rate while comparing to others algorithm. Especially, our simulation implies that FMC can reach better resource utilization while satisfying future service migration request. Keywords -- Network Function Virtualization, Service Function Chaining, Mobile Edge Computing