Elsevier

Journal of Systems and Software

Volume 143, September 2018, Pages 100-115
Journal of Systems and Software

A novel dynamic resource adjustment architecture for virtual tenant networks in SDN

https://doi.org/10.1016/j.jss.2018.04.033Get rights and content

Highlights

  • This work proposes a novel dynamic resource adjustment architecture.

  • The routing planning helps to solve the uneven distribution on routing paths.

  • The bandwidth resource planning helps to solve an excessive bandwidth requirement.

  • The proposed architecture helps to plan routing paths and guarantee bandwidth usage.

  • The proposed increases efficiency of routing assignment and bandwidth utilization.

Abstract

This work proposes a novel dynamic resource adjustment architecture that includes a routing planning mechanism and a bandwidth resource planning mechanism for a virtual tenant network (VTN) with software-defined networking (SDN). The routing planning mechanism helps to assign the routing paths in a physical network to solve the problem of their uneven distribution, with the ultimate goal of improving the utilization of the network routing. The bandwidth resource planning mechanism helps to manage the tenant demand for bandwidth resources to solve the problem of an excessive bandwidth requirement and thereby to improve the utilization of network bandwidth resources. The proposed architecture helps not only to plan routing paths in a physical network to satisfy a VTN user request but also to guarantee bandwidth usage based on overall network conditions. Analytical results indicate that the proposed routing planning mechanism increases the efficiency of routing assignment and that the proposed bandwidth resource scheduling increases bandwidth utilization by 10.92%.

Introduction

The proposed dynamic resource adjustment architecture is used for network virtualization in the OpenFlow network environment. The proposed routing planning mechanism and bandwidth resource planning mechanism in a virtual tenant network (VTN) improves network utilization and optimizes network resource usage. The routing planning mechanism assigns VTN bandwidth resources to all routing paths by calculating the costs of those paths. The bandwidth resource planning mechanism reassigns the resources when path utilization exceeds a threshold, to prevent bandwidth congestion.

Each VTN has a set of service-level agreements (SLAs), concerning bandwidth, delay time, routing path and other parameters. Fig. 1 shows bandwidth resource congestion in which the VTN routing path is assigned using the shortest path algorithm, and resource congestion arises from the assignment of all VTNs to one path. This work proposes routing planning mechanism to solve this problem.

When several users use routing path 2 at once, this path becomes overloaded. This work proposes a resource scheduling component of bandwidth resource planning to solve this problem. Fig. 2 shows a path overload scenario.

The rest of the paper is organized as follows. Section 2 presents the background to this study. Section 3 introduces the proposed dynamic resource adjustment architecture. Section 4 presents an analysis of the performance of the resource scheduling component. Finally, Section 5 draws conclusions.

Section snippets

Background knowledge

This section presents background knowledge that is related to software-defined networking (SDN) and network virtualization (Xiao et al., 2014, Betge-Brezetz et al., 2015, Li et al., 2016, Santos et al., 2014, Ortiz, 2013, Dixit et al., 2013). In this era of rapid growth of the Internet, the demand for network resources is increasing. Network diversity and complexity eliminate the usefulness of the traditional functions of network infrastructure for today's enterprizes, operators and end-users.

System overview

The routing planning mechanism and bandwidth resource planning mechanism that are proposed herein solve the problems of path congestion and bandwidth overload. Fig. 3 shows the dynamic resource adjustment architecture of system. The routing planning mechanism has two components - the SLA information collection component and the network policy component. The SLA information collection component collects and checks information about the SLA required by each VTN user, and sends messages to network

Performance analysis

This work proposes a test-bed with a controller, a coordinator, and a Mininet as the experimental environment. The controller, coordinator and Mininet run on Linux machines that communicate with each other using OpenFlow 1.3. To simulate a company network, the topology of the simulated network environment is similar to a tree, which is the structure that is frequently used by many companies. In the simulation, the maximum bandwidth capacity of each routing path is 100 Mbps. The total delay time

Conclusion

Network virtualization combines hardware and software network resources to be shared among users. A VTN application is an example of network virtualization. Static routing path policies may cause all VTN resources to be assigned to a single routing path. This work develops a routing planning mechanism and a bandwidth resource planning mechanism to prevent this outcome and a routing planning mechanism to assign VTN resources to routing paths as determined by path cost. To prevent the assignment

Acknowledgment

Authors thank for financial supports of the Ministry of Science and Technology, Taiwan under contract MOST 103-2221-E-011-069-MY2, 103-2221-E-011-067-MY2, 105-2221-E-011-078-MY3 and 105-2221-E-011-076-MY3.

Yi-Wei Ma is a lecturer in China Institute of FTZ Supply Chain, Shanghai Maritime University. He received the Ph.D. degree in Department of Engineering Science at National Cheng Kung University, Tainan, Taiwan. His research interests include internet of things, digital home network, cloud computing, embedded system, multimedia p2p streaming and ubiquitous computing.

References (34)

  • A. Furculita et al.

    Implementation issues for modified Dijkstra's and Floyd–Warshall algorithms in OpenFlow

  • M. Kalil et al.

    A framework for joint wireless network virtualization and cloud radio access networks for next generation wireless networks

    IEEE Access

    (2017)
  • M. Kim et al.

    Implementation of an OpenFlow network virtualization for multi-controller environment

  • E.J. Kitindi et al.

    Wireless network virtualization with SDN and C-RAN for 5 G networks: requirements, opportunities, and challenges

    IEEE Access

    (2017)
  • I. Kumara et al.

    Software-defined service networking: performance differentiation in shared multi-tenant cloud applications

    IEEE Trans. Serv. Comput.

    (2017)
  • L. Liao et al.

    Distributed FlowVisor: a distributed FlowVisor platform for quality of service aware cloud network virtualisation

    IET Netw.

    (2015)
  • A. Malishevskiy et al.

    OpenFlow-based network management with visualization of managed elements

  • Cited by (0)

    Yi-Wei Ma is a lecturer in China Institute of FTZ Supply Chain, Shanghai Maritime University. He received the Ph.D. degree in Department of Engineering Science at National Cheng Kung University, Tainan, Taiwan. His research interests include internet of things, digital home network, cloud computing, embedded system, multimedia p2p streaming and ubiquitous computing.

    Jiann-Liang Chen was born in Taiwan on December 15, 1963. He received the Ph.D. degree in Electrical Engineering from National Taiwan University, Taipei, Taiwan in 1989. Since August 2008, he has been with the Department of Electrical Engineering of National Taiwan University of Science and Technology, where he is a distinguished professor now. His current research interests are directed at cellular mobility management and personal communication systems.

    Chen-Chia Chang received the M.S. degree in Electrical Engineering of National Taiwan University of Science and Technology, Taipei, Taiwan. He research interest include SDN (Software-define networking), network virtualization and Virtual Tenant Networks.

    Akihiro Nakao is a professor at Applied Computer Science Course, Interfaculty Initiative in Information Studies, Graudate School of Interdisciplinary Information Studies, The University of Tokyo. He received Ph.D. degree in Computer Science from Princeton University. Aki has been leading NakaoLab at the University of Tokyo since 2005.

    Shu Yamamoto received the B.E, M.E, and Ph.D from the University of Tokyo, 1977, 1979, and 1989, respectively. He also received M.E from California institute of technology 1983. He joined KDDI (formerly KDD) in 1979. He is currently a project assistant professor of the University of Tokyo. He is currently engaged in Network Virtualization Research.

    View full text