Avaliação de desempenho de um controlador SDN implementado como uma VNF

In the present work, we performed experiments with an SDN controller implemented as a virtual network function in the KVM and XEN open source hypervisors and compared the most critical performance parameters in relation to a native scenario in order to measure the degradation of performance caused by Virtualization. We use Cbench to emulate SDN networks and evaluate the performance and latency of the Floodlight controller. We found that controller virtualization in the KVM environment resulted in degraded processing of flows by 29%, showed an increase in response time of 22% and utilized the processor much less than the Xen environment, which despite having overloaded the CPU in 25%, was able to achieve the best performance, response time and the lowest use of RAM. Resumo. No presente trabalho, realizamos experimentos com um controlador SDN implementado como uma função de rede virtual nos hypervisores de código aberto KVM e XEN e comparamos os parâmetros de desempenho mais críticos em relação a um cenário nativo com o propósito de mensurar a degradação de desempenho causada pela virtualização. Utilizamos Cbench para emular redes SDN e avaliar o desempenho e a latência do controlador Floodlight. Descobrimos que a virtualização do controlador no ambiente KVM resultou na degradação do processamento de fluxos em 29%, apresentou um aumento do tempo de resposta de 22% e utilizou bem menos o processador em relação ao ambiente Xen, que apesar de ter sobrecarregado a CPU em 25%, foi capaz de atingir o melhor desempenho, tempo de resposta e a menor utilização de memória RAM.

[1]  Rusty Russell,et al.  virtio: towards a de-facto standard for virtual I/O devices , 2008, OPSR.

[2]  Raja Elassali,et al.  A measurement of the response times of various OpenFlow/SDN controllers with CBench , 2015, 2015 IEEE/ACS 12th International Conference of Computer Systems and Applications (AICCSA).

[3]  Yang Xu,et al.  Toward High-Performance and Scalable Network Functions Virtualization , 2016, IEEE Internet Computing.

[4]  Subasish Mohapatra,et al.  Virtualization: A Survey on Concepts, Taxonomy and Associated Security Issues , 2010, 2010 Second International Conference on Computer and Network Technology.

[5]  Xiaolong Wu,et al.  Virtualization Technology and its Impact on Computer Hardware Architecture , 2011, 2011 Eighth International Conference on Information Technology: New Generations.

[6]  Ramakrishnan Rajamony,et al.  An updated performance comparison of virtual machines and Linux containers , 2015, 2015 IEEE International Symposium on Performance Analysis of Systems and Software (ISPASS).

[7]  Alexander Shalimov,et al.  Advanced study of SDN/OpenFlow controllers , 2013 .

[8]  Taekhee Kim,et al.  SDN and NFV benchmarking for performance and reliability , 2015, 2015 17th Asia-Pacific Network Operations and Management Symposium (APNOMS).

[9]  T. S. Eugene Ng,et al.  The Impact of Virtualization on Network Performance of Amazon EC2 Data Center , 2010, 2010 Proceedings IEEE INFOCOM.

[10]  Tiago Ferreto,et al.  Time Provisioning Evaluation of KVM, Docker and Unikernels in a Cloud Platform , 2016, 2016 16th IEEE/ACM International Symposium on Cluster, Cloud and Grid Computing (CCGrid).

[11]  Rastin Pries,et al.  A Flexible OpenFlow-Controller Benchmark , 2012, 2012 European Workshop on Software Defined Networking.

[12]  David Erickson,et al.  The beacon openflow controller , 2013, HotSDN '13.

[13]  F. Risso,et al.  Assessing the Performance of Virtualization Technologies for NFV: A Preliminary Benchmarking , 2015, 2015 Fourth European Workshop on Software Defined Networks.

[14]  Min Chen,et al.  Software-Defined Network Function Virtualization: A Survey , 2015, IEEE Access.

[15]  Marcelo G. Rubinstein,et al.  Proposing and evaluating the performance of a firewall implemented as a virtualized network function , 2016, 2016 7th International Conference on the Network of the Future (NOF).

[16]  Luigi Iannone,et al.  On the performance of SDN controllers: A reality check , 2015, 2015 IEEE Conference on Network Function Virtualization and Software Defined Network (NFV-SDN).

[17]  Marcelo G. Rubinstein,et al.  Performance evaluation of a virtualized HTTP proxy in KVM and Docker , 2016, 2016 7th International Conference on the Network of the Future (NOF).

[18]  Hongseok Jeon,et al.  Network service chaining challenges for VNF outsourcing in network function virtualization , 2015, 2015 International Conference on Information and Communication Technology Convergence (ICTC).

[19]  Lu Huang,et al.  Benchmarking Methodology for Virtualization Network Performance , 2017 .

[20]  A. Kivity,et al.  kvm : the Linux Virtual Machine Monitor , 2007 .

[21]  Filip De Turck,et al.  Network Function Virtualization: State-of-the-Art and Research Challenges , 2015, IEEE Communications Surveys & Tutorials.

[22]  Hongxin Hu,et al.  Performance considerations of network functions virtualization using containers , 2016, 2016 International Conference on Computing, Networking and Communications (ICNC).

[23]  Pavel Ivashchenko High performance in-kernel SDN / OpenFlow controller , 2014 .

[24]  Rob Sherwood,et al.  On Controller Performance in Software-Defined Networks , 2012, Hot-ICE.

[25]  Thierry Turletti,et al.  A Survey of Software-Defined Networking: Past, Present, and Future of Programmable Networks , 2014, IEEE Communications Surveys & Tutorials.

[26]  Muhammad Awais,et al.  Performance evaluation of OpenDaylight SDN controller , 2014, 2014 20th IEEE International Conference on Parallel and Distributed Systems (ICPADS).

[27]  Fabrice Bellard,et al.  QEMU, a Fast and Portable Dynamic Translator , 2005, USENIX ATC, FREENIX Track.

[28]  Mihai Barbulescu,et al.  A comparison of the performance and scalability of Xen and KVM hypervisors , 2013, 2013 RoEduNet International Conference 12th Edition: Networking in Education and Research.

[29]  Yong Yu,et al.  A Synthetical Performance Evaluation of OpenVZ, Xen and KVM , 2010, 2010 IEEE Asia-Pacific Services Computing Conference.