Link Capacity Estimation in Wireless Software Defined Networks
Download as PPTX, PDF2 likes575 views
The document discusses link capacity estimation in wireless software-defined networks (SDNs), outlining various methodologies for bandwidth and capacity estimation. It highlights the challenges associated with wireless mesh networks and presents the implementation of packet pair/train probing to estimate link capacity. The study concludes with experiments evaluating the effectiveness of this approach and suggests future work to validate findings in a real-world setting.
![Bandwidth and Capacity Estimation in
Traditional Networks (active)
• Variable Packet Size probing (VPS) [1], [2]
capacity of individual hops
• Self-Loading Periodic Streams (SLoPS) and Trains
of Packet Pairs (TOPP) [4], [5]
end-to-end available bandwidth
• Packet Pair/Train Dispersion probing (PPTD)[3]
end-to-end capacity of the path
Estimates
Estimate
Estimates
10](https://image.slidesharecdn.com/linkcapacity9-151125010428-lva1-app6892/85/Link-Capacity-Estimation-in-Wireless-Software-Defined-Networks-9-320.jpg)
![References
[1] S. M. Bellovin, “A best-case network performance model,” 1992.
[2] V. Jacobson, “Pathchar: A tool to infer characteristics of internet paths,” 1997.
[3] V. Jacobson, M. J. Karels, “Congestion avoidance and control,” in ACM SIGCOMM
computer communication review, vol. 18, no. 4. ACM, 1988, pp. 314–329.
[4] M. Jain and C. Dovrolis, End-to-end available bandwidth: Measurement methodology,
dynamics, and relation with TCP throughput. ACM, 2002, vol. 32, no. 4.
[5] B. Melander, M. Bj¨orkman, and P. Gunningberg, “A new end-to-end probing and
analysis method for estimating bandwidth bottlenecks,” in Global Telecommunications
Conference, 2000. GLOBECOM’00. IEEE, vol. 1. IEEE, 2000, pp. 415–420.
[6] J. Guerin, M. Portmann, K. Bialkowski, W. L. Tan, and S. Glass, “Lowcost wireless link
capacity estimation,” in Wireless Pervasive Computing (ISWPC), 2010 5th IEEE International
Symposium on. IEEE, 2010, pp. 343–348.
24](https://image.slidesharecdn.com/linkcapacity9-151125010428-lva1-app6892/85/Link-Capacity-Estimation-in-Wireless-Software-Defined-Networks-23-320.jpg)
Link Capacity Estimation in Wireless Software Defined Networks
- 1. Link Capacity Estimation in Wireless Software Defined Networks Farzaneh Pakzad, Marius Portmann, and Jared Hayward School of ITEE, The University of Queensland Brisbane, Australia Presented by Farzaneh Pakzad 25th International Telecommunication Networks and Applications Conference November 18-20, 2015, UNSW, Sydney, Australia 1
- 2. Overview • Background • Motivation • Bandwidth & Capacity Estimation in Traditional Networks • Packet Pair/Train Probing • Packet Pair/Train Probing in SDN • Experiments • Conclusion 2
- 3. 3 Background: Logical View of Software Defined Networking (SDN) Architecture
- 4. 4 Background: Logical View of Software Defined Networking (SDN) Architecture
- 5. The Promise of SDN 5
- 6. Background (contd.) Wireless Mesh Networks (WMNs) 6 Wireless Routers Gateways Mesh Clients Node Types Wireless Mesh Networks Challenges and Opportunities, Mihail L. Sichitiu, Electrical and Computer Eng. Dept., NC State University, Raleigh, NC, USA Public Safety Transportation Mining Enterprise Network Emergency Response
- 7. Potential of using SDN for WMNs 7 Limitations with WMNs: Network Topology Links Capacity
- 8. Available Bandwidth vs Link Capacity The maximum possible bandwidth of a link The maximum unused bandwidth 9
- 9. Bandwidth and Capacity Estimation in Traditional Networks (active) • Variable Packet Size probing (VPS) [1], [2] capacity of individual hops • Self-Loading Periodic Streams (SLoPS) and Trains of Packet Pairs (TOPP) [4], [5] end-to-end available bandwidth • Packet Pair/Train Dispersion probing (PPTD)[3] end-to-end capacity of the path Estimates Estimate Estimates 10
- 10. Packet Pair/Train Probing(Basic Idea) Sender Receiver Packet1 Packet2 Packet1Packet2 Packet 2 Packet 1 Back-to-Back Packet size: Link Capacity: Time Dispersion: LC C L 11
- 11. Packet Pair Probing in SDN 12 The implementation of Link capacity estimation mechanism on any Standard OpenFlow SDN controller SDN controller initiate sending of packet pair/train across a link
- 12. Packet Pair Probing in SDN Port 2 Port 1 13 Port 1 Port2 P1P2 Probing Packets P P1 P2 H1 H2 12 tt 1t2t LC d Set the “EtherType” of the Probe packet to unused value
- 13. Implementation/Experiments • Methodology – Considered simple topology described – Mininet: Linux based network emulator – Ns3: Emulate wireless links – Iperf: Measure wireless link capacity (as a reference) 14 • Ryu as our SDN controller platform
- 14. Link Capacity Estimation using Packet Pair Probing 15 Switch
- 15. Link Capacity Estimation using Packet Train Probing 40TLengthTrainPacket 16
- 16. Estimation Root mean Square Error(RMSE) and Overhead as a Function of Train Length (T) 17
- 17. Impact of Cross Traffic • Cross traffic cause underestimation of the link capacity • Two Type of Cross Traffic – Forward Cross Traffic Same Direction as the Probe Packets – Reverse Cross Traffic Reverse Direction of the Probe Packets 18
- 18. Impact of Cross Traffic 19 Distance d=0 Train Length T = 40 PPPPP Forward Cross Traffic Packet Train probe packets are interleaved with Reverse Cross Traffic
- 19. Compensate for the Impact of Reverse Cross Traffic • Controller query the port statistics from Switches , i.e. received packet count at port 2 of switch S1 20 Port 2 )( )1( )( T LRT TC R = the number of interleaved reverse cross traffic between the first and last packet of the train
- 20. Compensate for the Impact of Reverse Cross Traffic 21 Distance d=0 Train Length T = 40
- 21. Conclusions 22 Implemented a Prototype using Ryu SDN controller Good choice of Train length T Compensate the impact of cross traffic
- 22. Future work Evaluating this approach on a real wireless SDN test bed
- 23. References [1] S. M. Bellovin, “A best-case network performance model,” 1992. [2] V. Jacobson, “Pathchar: A tool to infer characteristics of internet paths,” 1997. [3] V. Jacobson, M. J. Karels, “Congestion avoidance and control,” in ACM SIGCOMM computer communication review, vol. 18, no. 4. ACM, 1988, pp. 314–329. [4] M. Jain and C. Dovrolis, End-to-end available bandwidth: Measurement methodology, dynamics, and relation with TCP throughput. ACM, 2002, vol. 32, no. 4. [5] B. Melander, M. Bj¨orkman, and P. Gunningberg, “A new end-to-end probing and analysis method for estimating bandwidth bottlenecks,” in Global Telecommunications Conference, 2000. GLOBECOM’00. IEEE, vol. 1. IEEE, 2000, pp. 415–420. [6] J. Guerin, M. Portmann, K. Bialkowski, W. L. Tan, and S. Glass, “Lowcost wireless link capacity estimation,” in Wireless Pervasive Computing (ISWPC), 2010 5th IEEE International Symposium on. IEEE, 2010, pp. 343–348. 24
- 24. 25