SlideShare a Scribd company logo
1 of 23
Distributed Mininet
with Symbiosis
Rong Rong, Jason Liu
Florida International University, Miami, Florida, USA
ICC’17, Paris, France, May 22, 2017
Supported by NSF CNS-1563883, DOD W911NF-13-1-0157, DOE LANL/LDRD, and USF/FC2 SEED
Outline
• Motivation
• Symbiotic Network Simulation
• Distributed Mininet with Symbiosis
• Experiments and Case Studies
• Conclusions
A Quick Look at Network Experiment Methods
3
WINLAB
ORBIT
• PhysicalTestbeds
• Real experiments
• Programmable
• Multiple users
• Relatively small
• Limited diversity
A Quick Look at Network Experiment Methods
4
GTNeTSSSFNet
And more …
• PhysicalTestbeds
• Simulation
• Flexible
• Scalable
• Diverse
• Lack credibility
• Significant development effort
A Quick Look at Network Experiment Methods
5
• PhysicalTestbeds
• Simulation
• Emulation
• Real apps or real operations
• Flexibility (from virtualization)
• Good but still limited in scale
• Flexible but limited scenarios
ModelNet
6
OpenVSwitch Linux namespace
Mininet is an emulation testbed for OpenFlow
Bob Lantz, Brandon Heller, and Nick McKeown. 2010. A net- work in a laptop: rapid prototyping
for software-defined networks. In Proceedings of the 9th ACM Workshop on Hot Topics in
7
OpenVSwitch Linux namespace
Distributed Mininet (Maxinet Approach)
P. Wette, M. Draxler,A. Schwabe, F. Wallaschek, M. Zahraee, and H. Karl, “Maxinet: Distributed emulation
of software-defined networks,” in Proceedings of the 2014 IFIP NetworkingConference, 2014, pp. 1–9.
Capture large-scale network behavior and
global traffic conditions
Simulation and Emulation
8
Network Simulation
• Full-scale network model
• Detailed topology and protocols
Network Emulation
• Real execution environment (operating
system, network stack, software tools)
• Unmodified applications
Represent real application traffic
behavior
Real-time Simulation
• Parallel Real-time Immersive network Modeling Environment (PRIME)
• Run parllel simulation in real time, interact with real networks
• Direct apps testing in large simulated network
• PrimoGENI: at-scale hybrid network experimentation on GENI
• Hybrid experiments with simulated,
emulated, and physical components
• Flexible configuration, real-time
visualization and steering
• May distribute across geographically
distributed sites
• Problem: tight coupling simulation
and emulation is bottleneck
9
Symbiotic Approach in a Nutshell
10
R2R1
H4
H2
H3
H1
SimulationEmulation
H1
H3
R1 R2
H2
H4
R3
R3
emulated flows
simulated flows
SimulationEmulation
effect of
simulated
flows
effect of
emulated
flows
Steady-State Queuing Model
11
Assuming M/D/1 Queue
Single-link Segment
Does ItWork?
• Well, steady-state does not work!
• Closed-form solution for transient effect is rather elusive, even
if assuming Poisson arrivals
• We invent a “control nob” to dynamically adjust μ* from
measurements
12
Adjust forTransient Effect
13
excess
service rate
avg pkt delay
in emulation
avg pkt delay
in simulation
excess queue
length
• The adjustment forces emulation to “track” the simulated network conditions
PreviousValidation Results
• Low, medium, heavy traffic conditions
• Various proportions of emulated vs. simulated traffic
• Mixed arrivals: exponential, constant, real traces
14
10% utilization 50% utilization 90% utilization
15
OpenVSwitch Linux namespace
Distributed Mininet with Symbiosis
Can we keep traffic within the same Mininet
instances if we can?
Distributed Mininet with Symbiosis
16
Practical Considerations
• Simulation Controller receives ”traffic demand“ from emulated
hosts (using strace) and reproduces the flows in simulation
accordingly (using implemented tcp variants)
• Simulation Controller receives periodic queue updates from
emulation (using traffic monitor): flow arrival rate, drop probability,
queuing delay
• Simulation Controller sends traffic control information to
emulation, which controls the real flows (using tc)
Simulation
Controller
Mininet
Instance
Mininet
Instance
Mininet
Instance
Mininet
Instance
Mininet
Instance
Mininet
Instance
Validation Experiment
18
Throughput(Mb/s)
A Case Study: Shrew Attack
• Attacker sends bursts of data at a regular interval to an over-committed bottleneck link[1]
• When burst intervals synchronize
with RTO ofTCP connections
sharing the link, they can trigger
TCP timeouts and consequently
strangle throughput
• Difficult to detect since avg.
traffic rate of the attack is low
19
[1] A. Kuzmanovic and E. W. Knightly, “Low-rateTCP-targeted denial of service attacks: The shrew
vs. the mice and elephants,” SIGCOMM 2003, pp. 75–86.
Victim
Attacker
Common
Bottleneck
Link
20
One “Good” Flow over 10 Mb/s Bottleneck Link
Distributed Mininet Results
21
Victim
Attacker
Common
Bottleneck
Link
Summary
• Emulation experiment is limited by scale
• Mininet has processing and memory limitation
• Maxinet can support large experiments, but has limitation on cross-
machine throughput
• Real-time simulation combines simulation and emulation, but is
also limited by the cross-system traffic
• Network Symbiosis decouples them and reduces cross-machine
communication (remove data traffic and keep control traffic)
• Symbiosis has limitations:
• Cross-system traffic is not allowed: need to combine with real-time method
• Dynamic traffic is not allowed: limited to studying long term traffic
22
ThankYou!

More Related Content

Similar to Distributed Mininet with Symbiosis

Virtual Time Machine for Large-Scale Reproducible Distributed Emulation
Virtual Time Machine for Large-Scale Reproducible Distributed EmulationVirtual Time Machine for Large-Scale Reproducible Distributed Emulation
Virtual Time Machine for Large-Scale Reproducible Distributed EmulationJason Liu
 
Cloud data management
Cloud data managementCloud data management
Cloud data managementambitlick
 
SAND: A Fault-Tolerant Streaming Architecture for Network Traffic Analytics
SAND: A Fault-Tolerant Streaming Architecture for Network Traffic AnalyticsSAND: A Fault-Tolerant Streaming Architecture for Network Traffic Analytics
SAND: A Fault-Tolerant Streaming Architecture for Network Traffic AnalyticsQin Liu
 
A data driven approach for monitoring network events
A data driven approach for monitoring network eventsA data driven approach for monitoring network events
A data driven approach for monitoring network eventsJisc
 
2005_NickTovar_Presentation
2005_NickTovar_Presentation2005_NickTovar_Presentation
2005_NickTovar_PresentationNicholas Tovar
 
RT15 Berkeley | ARTEMiS-SSN Features for Micro-grid / Renewable Energy Sourc...
RT15 Berkeley |  ARTEMiS-SSN Features for Micro-grid / Renewable Energy Sourc...RT15 Berkeley |  ARTEMiS-SSN Features for Micro-grid / Renewable Energy Sourc...
RT15 Berkeley | ARTEMiS-SSN Features for Micro-grid / Renewable Energy Sourc...OPAL-RT TECHNOLOGIES
 
PR-187 : MorphNet: Fast & Simple Resource-Constrained Structure Learning of D...
PR-187 : MorphNet: Fast & Simple Resource-Constrained Structure Learning of D...PR-187 : MorphNet: Fast & Simple Resource-Constrained Structure Learning of D...
PR-187 : MorphNet: Fast & Simple Resource-Constrained Structure Learning of D...Sunghoon Joo
 
Analytical Modeling of End-to-End Delay in OpenFlow Based Networks
Analytical Modeling of End-to-End Delay in OpenFlow Based NetworksAnalytical Modeling of End-to-End Delay in OpenFlow Based Networks
Analytical Modeling of End-to-End Delay in OpenFlow Based NetworksAzeem Iqbal
 
Improving the Efficiency of Cloud Infrastructures with Elastic Tandem Machine...
Improving the Efficiency of Cloud Infrastructures with Elastic Tandem Machine...Improving the Efficiency of Cloud Infrastructures with Elastic Tandem Machine...
Improving the Efficiency of Cloud Infrastructures with Elastic Tandem Machine...Frank Dürr
 
OpenKilda: Stream Processing Meets Openflow
OpenKilda: Stream Processing Meets OpenflowOpenKilda: Stream Processing Meets Openflow
OpenKilda: Stream Processing Meets OpenflowAPNIC
 
2009.08 grid peer-slides
2009.08 grid peer-slides2009.08 grid peer-slides
2009.08 grid peer-slidesYehia El-khatib
 
Real-time Classification of Malicious URLs on Twitter using Machine Activity ...
Real-time Classification of Malicious URLs on Twitter using Machine Activity ...Real-time Classification of Malicious URLs on Twitter using Machine Activity ...
Real-time Classification of Malicious URLs on Twitter using Machine Activity ...Pete Burnap
 
Real World Testbeds Emulation for Mobile Ad-hoc Networks
Real World Testbeds Emulation for Mobile Ad-hoc NetworksReal World Testbeds Emulation for Mobile Ad-hoc Networks
Real World Testbeds Emulation for Mobile Ad-hoc NetworksKishan Patel
 
chaos-monkey-increasing (1) (1)
chaos-monkey-increasing (1) (1)chaos-monkey-increasing (1) (1)
chaos-monkey-increasing (1) (1)Michael Alan Chang
 
Foundational Design Patterns for Multi-Purpose Applications
Foundational Design Patterns for Multi-Purpose ApplicationsFoundational Design Patterns for Multi-Purpose Applications
Foundational Design Patterns for Multi-Purpose ApplicationsChing-Hwa Yu
 
VeriFlow Presentation
VeriFlow PresentationVeriFlow Presentation
VeriFlow PresentationKrystle Bates
 
Composite Intrusion Detection in Process Control Networks
Composite Intrusion Detection in Process Control NetworksComposite Intrusion Detection in Process Control Networks
Composite Intrusion Detection in Process Control Networksguest8fdee6
 

Similar to Distributed Mininet with Symbiosis (20)

Virtual Time Machine for Large-Scale Reproducible Distributed Emulation
Virtual Time Machine for Large-Scale Reproducible Distributed EmulationVirtual Time Machine for Large-Scale Reproducible Distributed Emulation
Virtual Time Machine for Large-Scale Reproducible Distributed Emulation
 
Cloud data management
Cloud data managementCloud data management
Cloud data management
 
SAND: A Fault-Tolerant Streaming Architecture for Network Traffic Analytics
SAND: A Fault-Tolerant Streaming Architecture for Network Traffic AnalyticsSAND: A Fault-Tolerant Streaming Architecture for Network Traffic Analytics
SAND: A Fault-Tolerant Streaming Architecture for Network Traffic Analytics
 
A data driven approach for monitoring network events
A data driven approach for monitoring network eventsA data driven approach for monitoring network events
A data driven approach for monitoring network events
 
2005_NickTovar_Presentation
2005_NickTovar_Presentation2005_NickTovar_Presentation
2005_NickTovar_Presentation
 
RT15 Berkeley | ARTEMiS-SSN Features for Micro-grid / Renewable Energy Sourc...
RT15 Berkeley |  ARTEMiS-SSN Features for Micro-grid / Renewable Energy Sourc...RT15 Berkeley |  ARTEMiS-SSN Features for Micro-grid / Renewable Energy Sourc...
RT15 Berkeley | ARTEMiS-SSN Features for Micro-grid / Renewable Energy Sourc...
 
PR-187 : MorphNet: Fast & Simple Resource-Constrained Structure Learning of D...
PR-187 : MorphNet: Fast & Simple Resource-Constrained Structure Learning of D...PR-187 : MorphNet: Fast & Simple Resource-Constrained Structure Learning of D...
PR-187 : MorphNet: Fast & Simple Resource-Constrained Structure Learning of D...
 
Analytical Modeling of End-to-End Delay in OpenFlow Based Networks
Analytical Modeling of End-to-End Delay in OpenFlow Based NetworksAnalytical Modeling of End-to-End Delay in OpenFlow Based Networks
Analytical Modeling of End-to-End Delay in OpenFlow Based Networks
 
Lithium: Event-Driven Network Control
Lithium: Event-Driven Network ControlLithium: Event-Driven Network Control
Lithium: Event-Driven Network Control
 
Improving the Efficiency of Cloud Infrastructures with Elastic Tandem Machine...
Improving the Efficiency of Cloud Infrastructures with Elastic Tandem Machine...Improving the Efficiency of Cloud Infrastructures with Elastic Tandem Machine...
Improving the Efficiency of Cloud Infrastructures with Elastic Tandem Machine...
 
OpenKilda: Stream Processing Meets Openflow
OpenKilda: Stream Processing Meets OpenflowOpenKilda: Stream Processing Meets Openflow
OpenKilda: Stream Processing Meets Openflow
 
Introduction
IntroductionIntroduction
Introduction
 
2009.08 grid peer-slides
2009.08 grid peer-slides2009.08 grid peer-slides
2009.08 grid peer-slides
 
Real-time Classification of Malicious URLs on Twitter using Machine Activity ...
Real-time Classification of Malicious URLs on Twitter using Machine Activity ...Real-time Classification of Malicious URLs on Twitter using Machine Activity ...
Real-time Classification of Malicious URLs on Twitter using Machine Activity ...
 
Real World Testbeds Emulation for Mobile Ad-hoc Networks
Real World Testbeds Emulation for Mobile Ad-hoc NetworksReal World Testbeds Emulation for Mobile Ad-hoc Networks
Real World Testbeds Emulation for Mobile Ad-hoc Networks
 
chaos-monkey-increasing (1) (1)
chaos-monkey-increasing (1) (1)chaos-monkey-increasing (1) (1)
chaos-monkey-increasing (1) (1)
 
Foundational Design Patterns for Multi-Purpose Applications
Foundational Design Patterns for Multi-Purpose ApplicationsFoundational Design Patterns for Multi-Purpose Applications
Foundational Design Patterns for Multi-Purpose Applications
 
NS-3
NS-3 NS-3
NS-3
 
VeriFlow Presentation
VeriFlow PresentationVeriFlow Presentation
VeriFlow Presentation
 
Composite Intrusion Detection in Process Control Networks
Composite Intrusion Detection in Process Control NetworksComposite Intrusion Detection in Process Control Networks
Composite Intrusion Detection in Process Control Networks
 

Recently uploaded

Powerpoint exploring the locations used in television show Time Clash
Powerpoint exploring the locations used in television show Time ClashPowerpoint exploring the locations used in television show Time Clash
Powerpoint exploring the locations used in television show Time Clashcharlottematthew16
 
Beyond Boundaries: Leveraging No-Code Solutions for Industry Innovation
Beyond Boundaries: Leveraging No-Code Solutions for Industry InnovationBeyond Boundaries: Leveraging No-Code Solutions for Industry Innovation
Beyond Boundaries: Leveraging No-Code Solutions for Industry InnovationSafe Software
 
"Federated learning: out of reach no matter how close",Oleksandr Lapshyn
"Federated learning: out of reach no matter how close",Oleksandr Lapshyn"Federated learning: out of reach no matter how close",Oleksandr Lapshyn
"Federated learning: out of reach no matter how close",Oleksandr LapshynFwdays
 
Designing IA for AI - Information Architecture Conference 2024
Designing IA for AI - Information Architecture Conference 2024Designing IA for AI - Information Architecture Conference 2024
Designing IA for AI - Information Architecture Conference 2024Enterprise Knowledge
 
"ML in Production",Oleksandr Bagan
"ML in Production",Oleksandr Bagan"ML in Production",Oleksandr Bagan
"ML in Production",Oleksandr BaganFwdays
 
Vertex AI Gemini Prompt Engineering Tips
Vertex AI Gemini Prompt Engineering TipsVertex AI Gemini Prompt Engineering Tips
Vertex AI Gemini Prompt Engineering TipsMiki Katsuragi
 
Pigging Solutions in Pet Food Manufacturing
Pigging Solutions in Pet Food ManufacturingPigging Solutions in Pet Food Manufacturing
Pigging Solutions in Pet Food ManufacturingPigging Solutions
 
"LLMs for Python Engineers: Advanced Data Analysis and Semantic Kernel",Oleks...
"LLMs for Python Engineers: Advanced Data Analysis and Semantic Kernel",Oleks..."LLMs for Python Engineers: Advanced Data Analysis and Semantic Kernel",Oleks...
"LLMs for Python Engineers: Advanced Data Analysis and Semantic Kernel",Oleks...Fwdays
 
WordPress Websites for Engineers: Elevate Your Brand
WordPress Websites for Engineers: Elevate Your BrandWordPress Websites for Engineers: Elevate Your Brand
WordPress Websites for Engineers: Elevate Your Brandgvaughan
 
Unleash Your Potential - Namagunga Girls Coding Club
Unleash Your Potential - Namagunga Girls Coding ClubUnleash Your Potential - Namagunga Girls Coding Club
Unleash Your Potential - Namagunga Girls Coding ClubKalema Edgar
 
Story boards and shot lists for my a level piece
Story boards and shot lists for my a level pieceStory boards and shot lists for my a level piece
Story boards and shot lists for my a level piececharlottematthew16
 
Scanning the Internet for External Cloud Exposures via SSL Certs
Scanning the Internet for External Cloud Exposures via SSL CertsScanning the Internet for External Cloud Exposures via SSL Certs
Scanning the Internet for External Cloud Exposures via SSL CertsRizwan Syed
 
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...Integration and Automation in Practice: CI/CD in Mule Integration and Automat...
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...Patryk Bandurski
 
SIP trunking in Janus @ Kamailio World 2024
SIP trunking in Janus @ Kamailio World 2024SIP trunking in Janus @ Kamailio World 2024
SIP trunking in Janus @ Kamailio World 2024Lorenzo Miniero
 
costume and set research powerpoint presentation
costume and set research powerpoint presentationcostume and set research powerpoint presentation
costume and set research powerpoint presentationphoebematthew05
 
Artificial intelligence in cctv survelliance.pptx
Artificial intelligence in cctv survelliance.pptxArtificial intelligence in cctv survelliance.pptx
Artificial intelligence in cctv survelliance.pptxhariprasad279825
 
Gen AI in Business - Global Trends Report 2024.pdf
Gen AI in Business - Global Trends Report 2024.pdfGen AI in Business - Global Trends Report 2024.pdf
Gen AI in Business - Global Trends Report 2024.pdfAddepto
 
Advanced Test Driven-Development @ php[tek] 2024
Advanced Test Driven-Development @ php[tek] 2024Advanced Test Driven-Development @ php[tek] 2024
Advanced Test Driven-Development @ php[tek] 2024Scott Keck-Warren
 
Dev Dives: Streamline document processing with UiPath Studio Web
Dev Dives: Streamline document processing with UiPath Studio WebDev Dives: Streamline document processing with UiPath Studio Web
Dev Dives: Streamline document processing with UiPath Studio WebUiPathCommunity
 
SQL Database Design For Developers at php[tek] 2024
SQL Database Design For Developers at php[tek] 2024SQL Database Design For Developers at php[tek] 2024
SQL Database Design For Developers at php[tek] 2024Scott Keck-Warren
 

Recently uploaded (20)

Powerpoint exploring the locations used in television show Time Clash
Powerpoint exploring the locations used in television show Time ClashPowerpoint exploring the locations used in television show Time Clash
Powerpoint exploring the locations used in television show Time Clash
 
Beyond Boundaries: Leveraging No-Code Solutions for Industry Innovation
Beyond Boundaries: Leveraging No-Code Solutions for Industry InnovationBeyond Boundaries: Leveraging No-Code Solutions for Industry Innovation
Beyond Boundaries: Leveraging No-Code Solutions for Industry Innovation
 
"Federated learning: out of reach no matter how close",Oleksandr Lapshyn
"Federated learning: out of reach no matter how close",Oleksandr Lapshyn"Federated learning: out of reach no matter how close",Oleksandr Lapshyn
"Federated learning: out of reach no matter how close",Oleksandr Lapshyn
 
Designing IA for AI - Information Architecture Conference 2024
Designing IA for AI - Information Architecture Conference 2024Designing IA for AI - Information Architecture Conference 2024
Designing IA for AI - Information Architecture Conference 2024
 
"ML in Production",Oleksandr Bagan
"ML in Production",Oleksandr Bagan"ML in Production",Oleksandr Bagan
"ML in Production",Oleksandr Bagan
 
Vertex AI Gemini Prompt Engineering Tips
Vertex AI Gemini Prompt Engineering TipsVertex AI Gemini Prompt Engineering Tips
Vertex AI Gemini Prompt Engineering Tips
 
Pigging Solutions in Pet Food Manufacturing
Pigging Solutions in Pet Food ManufacturingPigging Solutions in Pet Food Manufacturing
Pigging Solutions in Pet Food Manufacturing
 
"LLMs for Python Engineers: Advanced Data Analysis and Semantic Kernel",Oleks...
"LLMs for Python Engineers: Advanced Data Analysis and Semantic Kernel",Oleks..."LLMs for Python Engineers: Advanced Data Analysis and Semantic Kernel",Oleks...
"LLMs for Python Engineers: Advanced Data Analysis and Semantic Kernel",Oleks...
 
WordPress Websites for Engineers: Elevate Your Brand
WordPress Websites for Engineers: Elevate Your BrandWordPress Websites for Engineers: Elevate Your Brand
WordPress Websites for Engineers: Elevate Your Brand
 
Unleash Your Potential - Namagunga Girls Coding Club
Unleash Your Potential - Namagunga Girls Coding ClubUnleash Your Potential - Namagunga Girls Coding Club
Unleash Your Potential - Namagunga Girls Coding Club
 
Story boards and shot lists for my a level piece
Story boards and shot lists for my a level pieceStory boards and shot lists for my a level piece
Story boards and shot lists for my a level piece
 
Scanning the Internet for External Cloud Exposures via SSL Certs
Scanning the Internet for External Cloud Exposures via SSL CertsScanning the Internet for External Cloud Exposures via SSL Certs
Scanning the Internet for External Cloud Exposures via SSL Certs
 
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...Integration and Automation in Practice: CI/CD in Mule Integration and Automat...
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...
 
SIP trunking in Janus @ Kamailio World 2024
SIP trunking in Janus @ Kamailio World 2024SIP trunking in Janus @ Kamailio World 2024
SIP trunking in Janus @ Kamailio World 2024
 
costume and set research powerpoint presentation
costume and set research powerpoint presentationcostume and set research powerpoint presentation
costume and set research powerpoint presentation
 
Artificial intelligence in cctv survelliance.pptx
Artificial intelligence in cctv survelliance.pptxArtificial intelligence in cctv survelliance.pptx
Artificial intelligence in cctv survelliance.pptx
 
Gen AI in Business - Global Trends Report 2024.pdf
Gen AI in Business - Global Trends Report 2024.pdfGen AI in Business - Global Trends Report 2024.pdf
Gen AI in Business - Global Trends Report 2024.pdf
 
Advanced Test Driven-Development @ php[tek] 2024
Advanced Test Driven-Development @ php[tek] 2024Advanced Test Driven-Development @ php[tek] 2024
Advanced Test Driven-Development @ php[tek] 2024
 
Dev Dives: Streamline document processing with UiPath Studio Web
Dev Dives: Streamline document processing with UiPath Studio WebDev Dives: Streamline document processing with UiPath Studio Web
Dev Dives: Streamline document processing with UiPath Studio Web
 
SQL Database Design For Developers at php[tek] 2024
SQL Database Design For Developers at php[tek] 2024SQL Database Design For Developers at php[tek] 2024
SQL Database Design For Developers at php[tek] 2024
 

Distributed Mininet with Symbiosis

  • 1. Distributed Mininet with Symbiosis Rong Rong, Jason Liu Florida International University, Miami, Florida, USA ICC’17, Paris, France, May 22, 2017 Supported by NSF CNS-1563883, DOD W911NF-13-1-0157, DOE LANL/LDRD, and USF/FC2 SEED
  • 2. Outline • Motivation • Symbiotic Network Simulation • Distributed Mininet with Symbiosis • Experiments and Case Studies • Conclusions
  • 3. A Quick Look at Network Experiment Methods 3 WINLAB ORBIT • PhysicalTestbeds • Real experiments • Programmable • Multiple users • Relatively small • Limited diversity
  • 4. A Quick Look at Network Experiment Methods 4 GTNeTSSSFNet And more … • PhysicalTestbeds • Simulation • Flexible • Scalable • Diverse • Lack credibility • Significant development effort
  • 5. A Quick Look at Network Experiment Methods 5 • PhysicalTestbeds • Simulation • Emulation • Real apps or real operations • Flexibility (from virtualization) • Good but still limited in scale • Flexible but limited scenarios ModelNet
  • 6. 6 OpenVSwitch Linux namespace Mininet is an emulation testbed for OpenFlow Bob Lantz, Brandon Heller, and Nick McKeown. 2010. A net- work in a laptop: rapid prototyping for software-defined networks. In Proceedings of the 9th ACM Workshop on Hot Topics in
  • 7. 7 OpenVSwitch Linux namespace Distributed Mininet (Maxinet Approach) P. Wette, M. Draxler,A. Schwabe, F. Wallaschek, M. Zahraee, and H. Karl, “Maxinet: Distributed emulation of software-defined networks,” in Proceedings of the 2014 IFIP NetworkingConference, 2014, pp. 1–9.
  • 8. Capture large-scale network behavior and global traffic conditions Simulation and Emulation 8 Network Simulation • Full-scale network model • Detailed topology and protocols Network Emulation • Real execution environment (operating system, network stack, software tools) • Unmodified applications Represent real application traffic behavior
  • 9. Real-time Simulation • Parallel Real-time Immersive network Modeling Environment (PRIME) • Run parllel simulation in real time, interact with real networks • Direct apps testing in large simulated network • PrimoGENI: at-scale hybrid network experimentation on GENI • Hybrid experiments with simulated, emulated, and physical components • Flexible configuration, real-time visualization and steering • May distribute across geographically distributed sites • Problem: tight coupling simulation and emulation is bottleneck 9
  • 10. Symbiotic Approach in a Nutshell 10 R2R1 H4 H2 H3 H1 SimulationEmulation H1 H3 R1 R2 H2 H4 R3 R3 emulated flows simulated flows SimulationEmulation effect of simulated flows effect of emulated flows
  • 11. Steady-State Queuing Model 11 Assuming M/D/1 Queue Single-link Segment
  • 12. Does ItWork? • Well, steady-state does not work! • Closed-form solution for transient effect is rather elusive, even if assuming Poisson arrivals • We invent a “control nob” to dynamically adjust μ* from measurements 12
  • 13. Adjust forTransient Effect 13 excess service rate avg pkt delay in emulation avg pkt delay in simulation excess queue length • The adjustment forces emulation to “track” the simulated network conditions
  • 14. PreviousValidation Results • Low, medium, heavy traffic conditions • Various proportions of emulated vs. simulated traffic • Mixed arrivals: exponential, constant, real traces 14 10% utilization 50% utilization 90% utilization
  • 15. 15 OpenVSwitch Linux namespace Distributed Mininet with Symbiosis Can we keep traffic within the same Mininet instances if we can?
  • 16. Distributed Mininet with Symbiosis 16
  • 17. Practical Considerations • Simulation Controller receives ”traffic demand“ from emulated hosts (using strace) and reproduces the flows in simulation accordingly (using implemented tcp variants) • Simulation Controller receives periodic queue updates from emulation (using traffic monitor): flow arrival rate, drop probability, queuing delay • Simulation Controller sends traffic control information to emulation, which controls the real flows (using tc) Simulation Controller Mininet Instance Mininet Instance Mininet Instance Mininet Instance Mininet Instance Mininet Instance
  • 19. A Case Study: Shrew Attack • Attacker sends bursts of data at a regular interval to an over-committed bottleneck link[1] • When burst intervals synchronize with RTO ofTCP connections sharing the link, they can trigger TCP timeouts and consequently strangle throughput • Difficult to detect since avg. traffic rate of the attack is low 19 [1] A. Kuzmanovic and E. W. Knightly, “Low-rateTCP-targeted denial of service attacks: The shrew vs. the mice and elephants,” SIGCOMM 2003, pp. 75–86. Victim Attacker Common Bottleneck Link
  • 20. 20 One “Good” Flow over 10 Mb/s Bottleneck Link
  • 22. Summary • Emulation experiment is limited by scale • Mininet has processing and memory limitation • Maxinet can support large experiments, but has limitation on cross- machine throughput • Real-time simulation combines simulation and emulation, but is also limited by the cross-system traffic • Network Symbiosis decouples them and reduces cross-machine communication (remove data traffic and keep control traffic) • Symbiosis has limitations: • Cross-system traffic is not allowed: need to combine with real-time method • Dynamic traffic is not allowed: limited to studying long term traffic 22

Editor's Notes

  1. Some parallelized
  2. Planetlab has ~1000 nodes, but shared among many users Fixed setup
  3. Symbiosis infers mutual beneficial relations between two entities Real application testing under large-sale diverse network conditions
  4. Emulated hosts refer to machines running real systems and real applications Emulated traffic refers to real network traffic, supposedly flowing on the simulated network Real traffic needs to zip through segments, intermixing with simulated traffic