Using hardware classification to improve PC-based OpenFlow switching

OpenFlow is a promising technology that offers a flexible flow-based forwarding scheme based on multiple fields in the packet headers of different protocol layers. It provides a feasibility to adopt new network protocols and innovations thanks to a more flexible forwarding compared to traditional IP forwarding. In this paper, we propose an architectural design to improve lookup performance of PC-based OpenFlow switching in Linux using a standard commodity network interface card based on the Intel 82599 Gigabit Ethernet controller. We introduce a fast data path based on caching of flow table entries in on-board classification hardware on the NIC. We describe our design, and we present an experimental evaluation of its performance. Our results show that the throughput of our fast path gives over 40 percent higher throughput compared to the regular software-based OpenFlow switching.

[1]  George Varghese,et al.  New directions in traffic measurement and accounting: Focusing on the elephants, ignoring the mice , 2003, TOCS.

[2]  Kun-Chan Lan,et al.  A measurement study of correlations of Internet flow characteristics , 2006, Comput. Networks.

[3]  Ramakrishna Kotla,et al.  Proceedings of the first ACM asia-pacific workshop on Workshop on systems , 2010, SIGCOMM 2010.

[4]  Nevil Brownlee.  Some Observations of Internet Stream Lifetimes , 2005, PAM.

[5]  Markus Hidell,et al.  Improving PC-based OpenFlow switching performance , 2010, 2010 ACM/IEEE Symposium on Architectures for Networking and Communications Systems (ANCS).

[6]  Scott O. Bradner,et al.  Benchmarking Methodology for Network Interconnect Devices , 1996, RFC.

[7]  R. Wilder,et al.  Wide-area Internet traffic patterns and characteristics , 1997, IEEE Netw..

[8]  Anirban Mahanti,et al.  Byte me: a case for byte accuracy in traffic classification , 2007, MineNet '07.

[9]  Andrea Bianco,et al.  OpenFlow Switching: Data Plane Performance , 2010, 2010 IEEE International Conference on Communications.

[10]  Julio Ortega Lopera,et al.  Accelerating OpenFlow switching with network processors , 2009, ANCS '09.

[11]  Daniel Turull Torrents.  Open source traffic analyzer , 2010 .

[12]  Cristina Cervello-Pastor,et al.  With evolution for revolution: managing FEDERICA for future Internet research , 2009, IEEE Communications Magazine.

[13]  Hiroaki Harai.  Designing new-generation network - overview of AKARI Architecture Design - , 2009, 2009 Asia Communications and Photonics conference and Exhibition (ACP).

[14]  David A. Koufaty,et al.  Hyperthreading Technology in the Netburst Microarchitecture , 2003, IEEE Micro.

[15]  Jonathan S. Turner.  A proposed architecture for the GENI backbone platform , 2006, 2006 Symposium on Architecture For Networking And Communications Systems.

[16]  Anja Feldmann,et al.  Dagstuhl perspectives workshop on end-to-end protocols for the future internet , 2009, CCRV.

[17]  Kok-Kiong Yap,et al.  Towards software-friendly networks , 2010, APSys '10.

[18]  D. Marr,et al.  Hyper-Threading Technology Architecture and MIcroarchitecture , 2002 .

[19]  Guido Appenzeller,et al.  Implementing an OpenFlow switch on the NetFPGA platform , 2008, ANCS '08.

[20]  Reza Rooholamini,et al.  An Empirical Study of Hyper-Threading in High-Performance Computing Clusters , 2002 .

[21]  Raouf Boutaba,et al.  Network virtualization: state of the art and research challenges , 2009, IEEE Communications Magazine.

[22]  Nick McKeown,et al.  OpenFlow: enabling innovation in campus networks , 2008, CCRV.

[23]  kc claffy,et al.  Understanding Internet traffic streams: dragonflies and tortoises , 2002, IEEE Commun. Mag..