Floodlight overview & performance comparison by patrick huang
Download as PPTX, PDF•4 likes•3,159 views
This document compares reactive and proactive approaches for OpenFlow controllers and discusses their scalability. It summarizes the results of benchmarking different OpenFlow controllers. Reactive controllers like Floodlight and NOX see round-trip times increase rapidly with more switches due to weak hardware. Maestro has more steady performance. Reactive approaches have efficient flow table usage but slow setup times, while proactive approaches have zero setup time but hard management. A hybrid approach may improve efficiency and scalability.
Floodlight overview & performance comparison by patrick huang
- 1. Floodlight Overview & Performance Comparison Speaker: Patrick Huang Advisor: H.P. Wen
- 2. Outline 0 Flashlight Overview 0 A Flexible OpenFlow-Controller Benchmark 0 Comparing OpenFlow Controller Paradigms Scalability: Reactive and Proactive
- 3. Outline 0 Flashlight Overview 0 Reference list: 0 Project Floodlight, http://www.projectfloodlight.org/floodlight/ 0 Slideshare 0Big Switch 0 A Flexible OpenFlow-Controller Benchmark 0 Comparing OpenFlow Controller Paradigms Scalability: Reactive and Proactive
- 4. Why Floodlight? 0 Openflow 0 Work with phy- and vir- switches that speak openflow protocol 0 Apache Licensed 0 Use floodlight for any purpose 0 Open community 0 Developed by open community 0 Easy to use 0 Floodlight is drop dead simple to build and run. 0 enterprise-class
- 5. OpenFlow implemented by Floodlight0 Application Tier 0 Control Plane Tier 0 Data plane Tier 0 Indigo Data Plane Interface 0 an open source project 0 aimed at enabling support for OpenFlow on physical and hypervisor switches.
- 6. OpenFlow implemented by Floodlight 0 Floodlight 0 An Openflow controller ,and 0 A collection of applications built on top the floodlight controller 0 Applications on top of it 0 solve different user needs over the network
- 7. OpenFlow implemented by Floodlight0 Flood light controller 0 Core service of common interest to SDN applications 0 Module Application 0 Applications with higher bandwidth communication with controller 0 REST Application 0 Application in any language leveraging service via REST API exposed by controller modules and module applicaitons
- 8. Floodlight Architecture Topology - Tracks links between hosts and switches Device Manager - Tracks devices in the network (MACs, IPs, etc.) Storage - Abstraction layer for storing controller storage. Memory is used. Counter Store – Openflow + Floodlight stats Routing / Forwarding – Core engine for storing, calculating paths and installing flows. Web UI- Rest APIs Topology Device Manager Learning Switch Hub Web Storage Counter Store Floodlight Controller Routing / Forward Static FlowPusher REST APIs Module App
- 9. Floodlight Architecture Learning Switch - Can replace Routing / Forwarding(v) Hub - Can replace routing / forwarding(x) Main difference: REST API provider Topology Device Manager Learning Switch Hub Web Storage Counter Store Floodlight Controller Routing / Forward Static FlowPusher REST APIs Module App
- 10. Open Controllers 11 Name Lang Platform (s) License Original Author Notes OpenFlow Reference C Linux OpenFlow License Stanford/Nicir a not designed for extensibility NOX Python, C++ Linux GPL Nicira actively developed Beacon Java Win, Mac, Linux, Android GPL (core), FOSS Licenses for your code David Erickson (Stanford) runtime modular, web UI framework, regression test framework Maestro Java Win, Mac, Linux LGPL Zheng Cai (Rice) Trema Ruby, C Linux GPL NEC includes emulator, regression test framework Floodlight Java Win, Mac, Linux Apache Big Switch Apache licensed, actively developed
- 11. Learning More? 0 Check out the website: 0 http://floodlight.openflowhub.org 0 Join the mailing list: 0 http://groups.google.com/a/openflowhub.org/group/fl oodlight-dev/topics 0 Get the code: 0 http://floodlight.openflowhub.org/display/Floodlight/F loodlight+Downloads
- 12. Outline 0 Flashlight Overview 0 A Flexible OpenFlow-Controller Benchmark 0 Soft Defined Networking(EWSDN), 2012 Euro Workshop on 0 Univ. of Wurzburg, Wurzburg, Germany 0 Comparing OpenFlow Controller Paradigms Scalability: Reactive and Proactive
- 13. A Flexible OpenFlow-Controller Benchmark 0 Goal 0 Introduce a tool to achieve a flexible Openflow Controller benchmark 0 Methods 0 The Benchmark creates a set of message-generating virtual switches
- 14. Related work 0 Measure OF switching performance 0 OF switches not designed as flow switches Often performance bottlenecks 0 Cbench
- 15. Result: Mean Round Trip Time 0 RTT: Interval of 0 Packet-In message is dispatched from the virtual switch to the controller 0 Packet-Out or FlowMod message is received by the switch.
- 16. Result: RTT 0 Flood light and Nox Behave similarly 0 Increase rapidly from 0 200ms, 1 switch 0 6 seconds, 30 switches 0 Both controllers are under heavy load at the point 0 Due to weak hardware
- 17. Result: RTT 0 Maestro start at RTT 6ms 0 Larger the no. of switches, larger RTT 0 But, far steady than two others
- 18. Result: Send and ReceptionRate 0 The rate is accepts packets 0 Provide insights into rate control and polling strategies 0 No. of packets sent from switches to the controller
- 19. Virtual Switch Packet-in Send- rate0 Send rate: From switches to controller through OF secure channel 0 Floodlight 0 do not increase obviously with the no. of switches 0 Start at 10000 pps, one switch 0 Increase to 38000 pps 0 NOX 0 10000-70000 0 Maestro 0 5000-140000
- 20. Virtual Switch Packet-in Send- rate0 Suggestion: 0 NOX and Floodlight 0 Implementation of rate control mechanism 0 Maestro 0 Accept packet in a best effort manner
- 21. Virtual Switch Packet-out Reception- rate 0 The no. of responses the switches recieves 0 No increase for floodlight and NOX 0 Stable at 10000 pps 0 Maestro 0 From 5000-135000 0 Outstanding packets 0 No. of unanserwed messages by controller
- 22. Virtual Switch Packet-out Reception- rate 0 Outstanding packets 0 No. of unanserwed messages by controller 0 140000-135000
- 23. Outline 0 Flashlight Overview 0 A Flexible OpenFlow-Controller Benchmark 0 Comparing OpenFlow Controller Paradigms Scalability: Reactive and Proactive 0 Advanced Information Networking and Applications (AINA), 2013 0 Univ. Estadual do Ceara (UECE), Fortaleza, Brazil 0 Citation: 1
- 24. Comparing OpenFlow Controller Paradigms Scalability: Reactive and Proactive 0 Analysis 0 Demo the flaw of reactive approach 0 Conclusion 0 Indicate the effectiveness of a hybrid approach to improve the efficiency and scalability of OF architecture
- 25. Introduction 0 Issue: Scalability 0 Only one controller 0 As the no. of OpenFlow switches increases 0 Control messages to the centralized controller grows 0 With increase of network diameter, switches have longer setup delay 0 Bounded by controller’s processor power
- 26. Related Work 0 Hyperfolw: 0 To provide scalability Use as many controllers as necessary But keep network control logically centralized 0 DevoFlow, to provide scalability 0 devolve network control to switch 0 Introduce two new mechanisms to be imple. On swtich Rule cloning Local actions
- 27. Related Work • Source-Flow, to reduce no. of flow entries – Try to reduce no. of flow entries • Use MPLS-like tunneling approach to reduce Ternary Content Addressable Memory used space
- 28. OF architecture • Reactive – First packets of flow triggers controller to insert flow table – Pros efficient use of flow table memory – Cons Cause setup time Hard dependency, connection must retain
- 29. OF architecture • Proactive – Controller pre-populate flow table in switches – Pros zero setup time Soft dependency – Cons Hard management
- 31. Evaluating OpenFlow controller’s Parafigm • Run Cbench to stress the controller’s capacity • The benchmarking measurement – flows per sec that can be processed by controller
- 33. Thank You
Editor's Notes
- #12: Language open-sourcePerformance cross-platform runtime modular