FlowVisor‐based cost‐aware VN embedding in OpenFlow networks. (4th July 2016)

• Record Type: Journal Article
• Authors: Zhong, Xuxia; Wang, Ying; Qiu, Xuesong; Li, Wenjing
• Other Names: Meirosu Catalin guestEditor.; Pentikousis Kostas guestEditor.; Kind Mario guestEditor.; Gonzalez Prieto Alberto guestEditor.
• Is Part Of: International journal of network management. Volume 26:Number 5(2016:Sep./Oct.)
• Page Start: 373
• Page End: 395
• Publication Date: 2016-07-04
• Subjects: Computer networks -- Management -- Periodicals; 004.6
• DOI: 10.1002/nem.1940
• Languages: English
• ISSNs: 1055-7148

Abstract

Summary: Software‐defined network has emerged as a promising network technology to simplify the network management and enable innovation through network programmability. With the help of FlowVisor, network virtualization is achieved in software‐defined networks. We mainly focus on investigating how to embed virtual network requests with different network sizes (small, regular, and large) in a cost‐efficient way. However, the features of FlowVisor and limitations of OpenFlow switches, such as the presence of a central controller, the limited flow space resource, and the out‐of‐band bandwidth requirement, bring new challenges. In this paper, we first formulate the FlowVisor‐enabled cost‐aware virtual network embedding issue as an integer linear programming problem. Then, we propose two embedding algorithms: a heuristic algorithm LF‐OVNE and a meta‐heuristic algorithm CA‐CVNC‐EPSO. Meanwhile, a separate step is taken to select a cost‐aware controller for each slice. Finally, we evaluate our algorithms and compare the embedding performance with the existing algorithms. The simulation results show that our algorithms perform the online virtual network embedding in a more cost‐efficient way. Copyright © 2016 John Wiley & Sons, Ltd. Abstract : We have explored the virtual network embedding techniques in OpenFlow‐enabled networks with the help of FlowVisor. FlowVisor is used to slice the substrate network resources according to users' requirements. We take into account flow entry constraint of OpenFlow switches, controller placement of slices and bandwidth consumption of control network. Two algorithms have been proposed to accomplish the cost‐aware embedding problem.