Switch-visor: towards infrastructure-level virtualization of SDN switches

To test and update switch operating systems, developers and testers need to install run beta-switch OSes (switch agents) alongside production versions. However, today's network virtualization solutions fail to support infrastructure-level virtualization of hardware switches. In particular, they fail to provide performance guarantee and isolation of the switch's resources: CPU, Memory, and ASIC (TCAM/SRAM). In this paper, we define the notion of infrastructure-level switch virtualization, akin to IaaS, infrastructure-level switch virtualization provides tenants, testers or developers, with low-level control over the switches: allowing a tenant to install switch agents on the switches and to run their own controller. To support this abstraction, we present a system, Switch-Visor, which presents a first step towards providing comprehensive virtualization of a switch's resources. Switch-Visor employs a synthesis of well-founded virtualization technologies and novel hardware virtualization techniques. Switch-Visor introduces three main concepts: first, using container-based virtualization on the switch to virtualize CPU and Memory; second, leveraging intelligent TCAM management and novel schedulers to provide guarantees within the ASIC, and employing novel domain-specific offloading techniques to eliminate sources of interference. Our proposed solutions, leverage changes to switch OS and switch agents making them immediately applicable to existing SDN switches.

[1]  Laurent Vanbever,et al.  HotSwap: correct and efficient controller upgrades for software-defined networks , 2013, HotSDN '13.

[2]  Marina Thottan,et al.  Mazu: Taming Latency in Software Defined Networks , 2014 .

[3]  Rob Sherwood,et al.  Can the Production Network Be the Testbed? , 2010, OSDI.

[4]  R. Weisberg.  A-N-D , 2011 .

[5]  Wolfgang Kellerer,et al.  hvbench: An open and scalable SDN network hypervisor benchmark , 2016, 2016 IEEE NetSoft Conference and Workshops (NetSoft).

[6]  Sujata Banerjee,et al.  DevoFlow: scaling flow management for high-performance networks , 2011, SIGCOMM.

[7]  Ye Wang,et al.  Shadow configuration as a network management primitive , 2008, SIGCOMM '08.

[8]  William E. Weihl,et al.  Lottery scheduling: flexible proportional-share resource management , 1994, OSDI '94.

[9]  Minlan Yu,et al.  Virtually eliminating router bugs , 2009, CoNEXT '09.

[10]  Benjamin Hindman,et al.  Dominant Resource Fairness: Fair Allocation of Multiple Resource Types , 2011, NSDI.

[11]  Scott Shenker,et al.  Analysis and simulation of a fair queueing algorithm , 1989, SIGCOMM '89.

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

[13]  Xin Huang,et al.  Tango: Simplifying SDN Control with Automatic Switch Property Inference, Abstraction, and Optimization , 2014, CoNEXT.

[14]  Liming Chen,et al.  N-VERSION PROGRAMMINC: A FAULT-TOLERANCE APPROACH TO RELlABlLlTY OF SOFTWARE OPERATlON , 1995, Twenty-Fifth International Symposium on Fault-Tolerant Computing, 1995, ' Highlights from Twenty-Five Years'..

[15]  Mathieu Bouet,et al.  Improving SDN with InSPired Switches , 2016, SOSR.

[16]  Adam J. Aviv,et al.  Enabling Practical Software-defined Networking Security Applications with OFX , 2016, NDSS.

[17]  Xin Wu,et al.  FOCUS: Function Offloading from a Controller to Utilize Switch power , 2016, 2016 IEEE Conference on Network Function Virtualization and Software Defined Networks (NFV-SDN).

[18]  Brian D. Noble,et al.  Bobtail: Avoiding Long Tails in the Cloud , 2013, NSDI.

[19]  Theophilus Benson,et al.  The Case for Making Tight Control Plane Latency Guarantees in SDN Switches , 2017, SOSR.

[20]  Jialin Li,et al.  Tales of the Tail: Hardware, OS, and Application-level Sources of Tail Latency , 2014, SoCC.

[21]  Guru M. Parulkar,et al.  OpenVirteX: make your virtual SDNs programmable , 2014, HotSDN.

[22]  Fang Hao,et al.  Scotch: Elastically Scaling up SDN Control-Plane using vSwitch based Overlay , 2014, CoNEXT.