1 MaxNet and TCP Reno/RED on mice traffic Khoa Truong Phan Ho Chi Minh city University of Technology (HCMUT)

Slides:



Advertisements
Similar presentations
Martin Suchara, Ryan Witt, Bartek Wydrowski California Institute of Technology Pasadena, U.S.A. TCP MaxNet Implementation and Experiments on the WAN in.
Advertisements

Congestion Control and Fairness Models Nick Feamster CS 4251 Computer Networking II Spring 2008.
FAST TCP Anwis Das Ajay Gulati Slides adapted from : IETF presentation slides Link:
Cheng Jin David Wei Steven Low FAST TCP: design and experiments.
TCP Vegas: New Techniques for Congestion Detection and Control.
Congestion Control Created by M Bateman, A Ruddle & C Allison As part of the TCP View project.
CS640: Introduction to Computer Networks Mozafar Bag-Mohammadi Lecture 3 TCP Congestion Control.
Congestion Control: TCP & DC-TCP Swarun Kumar With Slides From: Prof. Katabi, Alizadeh et al.
1 End to End Bandwidth Estimation in TCP to improve Wireless Link Utilization S. Mascolo, A.Grieco, G.Pau, M.Gerla, C.Casetti Presented by Abhijit Pandey.
Advanced Computer Networking Congestion Control for High Bandwidth-Delay Product Environments (XCP Algorithm) 1.
The War Between Mice and Elephants LIANG GUO, IBRAHIM MATTA Computer Science Department Boston University ICNP (International Conference on Network Protocols)
Congestion Control An Overview -Jyothi Guntaka. Congestion  What is congestion ?  The aggregate demand for network resources exceeds the available capacity.
XCP: Congestion Control for High Bandwidth-Delay Product Network Dina Katabi, Mark Handley and Charlie Rohrs Presented by Ao-Jan Su.
Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 1 ECSE-6600: Internet Protocols Informal Quiz #07 Shivkumar Kalyanaraman: GOOGLE: “Shiv RPI”
1 USC INFORMATION SCIENCES INSTITUTE RAP: An End-to-End Congestion Control Mechanism for Realtime Streams in the Internet Reza Rejaie, Mark Handley, Deborah.
The War Between Mice and Elephants Presented By Eric Wang Liang Guo and Ibrahim Matta Boston University ICNP
AQM for Congestion Control1 A Study of Active Queue Management for Congestion Control Victor Firoiu Marty Borden.
15-744: Computer Networking L-10 Congestion Control.
Cheng Jin David Wei Steven Low FAST TCP: Motivation, Architecture, Algorithms, Performance.
1 Random Early Detection Gateways for Congestion Avoidance Sally Floyd and Van Jacobson, IEEE Transactions on Networking, Vol.1, No. 4, (Aug 1993), pp
Transport: TCP Manpreet Singh (Slides borrowed from various sources on the web)
1 Emulating AQM from End Hosts Presenters: Syed Zaidi Ivor Rodrigues.
Random Early Detection Gateways for Congestion Avoidance
Comparison of MaxNet and XCP: Network Congestion Control using explicit signalling Speaker: Bartek Wydrowski Compiled from work by: Lachlan Andrew (2),
TCP Congestion Control
Congestion Control for High Bandwidth-delay Product Networks Dina Katabi, Mark Handley, Charlie Rohrs.
1 A State Feedback Control Approach to Stabilizing Queues for ECN- Enabled TCP Connections Yuan Gao and Jennifer Hou IEEE INFOCOM 2003, San Francisco,
Adaptive Control for TCP Flow Control Thesis Presentation Amir Maor.
Advanced Computer Networks : RED 1 Random Early Detection Gateways for Congestion Avoidance Sally Floyd and Van Jacobson, IEEE Transactions on Networking,
Introduction 1 Lecture 14 Transport Layer (Congestion Control) slides are modified from J. Kurose & K. Ross University of Nevada – Reno Computer Science.
Transport Layer 4 2: Transport Layer 4.
CS144 An Introduction to Computer Networks
CS540/TE630 Computer Network Architecture Spring 2009 Tu/Th 10:30am-Noon Sue Moon.
Understanding the Performance of TCP Pacing Amit Aggarwal, Stefan Savage, Thomas Anderson Department of Computer Science and Engineering University of.
Link Scheduling & Queuing COS 461: Computer Networks
ACN: RED paper1 Random Early Detection Gateways for Congestion Avoidance Sally Floyd and Van Jacobson, IEEE Transactions on Networking, Vol.1, No. 4, (Aug.
Parameswaran, Subramanian
MaxNet NetLab Presentation Hailey Lam Outline MaxNet as an alternative to TCP Linux implementation of MaxNet Demonstration of fairness, quick.
27th, Nov 2001 GLOBECOM /16 Analysis of Dynamic Behaviors of Many TCP Connections Sharing Tail-Drop / RED Routers Go Hasegawa Osaka University, Japan.
High-speed TCP  FAST TCP: motivation, architecture, algorithms, performance (by Cheng Jin, David X. Wei and Steven H. Low)  Modifying TCP's Congestion.
Acknowledgments S. Athuraliya, D. Lapsley, V. Li, Q. Yin (UMelb) S. Adlakha (UCLA), J. Doyle (Caltech), K. Kim (SNU/Caltech), F. Paganini (UCLA), J. Wang.
Analysis of Buffer Size in Core Routers by Arthur Dick Supervisor Anirban Mahanti.
What is TCP? Connection-oriented reliable transfer Stream paradigm
T. S. Eugene Ngeugeneng at cs.rice.edu Rice University1 COMP/ELEC 429/556 Introduction to Computer Networks Principles of Congestion Control Some slides.
Jennifer Rexford Fall 2014 (TTh 3:00-4:20 in CS 105) COS 561: Advanced Computer Networks TCP.
 Last Class  This Class  Chapter 6.3. ~ 6.4.  TCP congestion control.
1 Analysis of a window-based flow control mechanism based on TCP Vegas in heterogeneous network environment Hiroyuki Ohsaki Cybermedia Center, Osaka University,
David Wetherall Professor of Computer Science & Engineering Introduction to Computer Networks Fairness of Bandwidth Allocation (§6.3.1)
We used ns-2 network simulator [5] to evaluate RED-DT and compare its performance to RED [1], FRED [2], LQD [3], and CHOKe [4]. All simulation scenarios.
The Macroscopic behavior of the TCP Congestion Avoidance Algorithm.
TCP Westwood: Efficient Transport for High-speed wired/wireless Networks 2008.
TCP transfers over high latency/bandwidth networks & Grid DT Measurements session PFLDnet February 3- 4, 2003 CERN, Geneva, Switzerland Sylvain Ravot
1 Advanced Transport Protocol Design Nguyen Multimedia Communications Laboratory March 23, 2005.
Spring Computer Networks1 Congestion Control Sections 6.1 – 6.4 Outline Preliminaries Queuing Discipline Reacting to Congestion Avoiding Congestion.
© Janice Regan, CMPT 128, CMPT 371 Data Communications and Networking Congestion Control 0.
Peer-to-Peer Networks 13 Internet – The Underlay Network
ECEN 619, Internet Protocols and Modeling Prof. Xi Zhang Random Early Detection Gateways for Congestion Avoidance Sally Floyd and Van Jacobson, IEEE Transactions.
Congestion Avoidance Created by M Bateman, A Ruddle & C Allison As part of the TCP View project.
Window Control Adjust transmission rate by changing Window Size
COMP 431 Internet Services & Protocols
Introduction to Congestion Control
TCP Vegas: New Techniques for Congestion Detection and Avoidance
TCP Congestion Control
TCP Congestion Control
So far, On the networking side, we looked at mechanisms to links hosts using direct linked networks and then forming a network of these networks. We introduced.
The University of Adelaide, School of Computer Science
TCP Congestion Control
If both sources send full windows, we may get congestion collapse
TCP Congestion Control
Computer Science Division
Presentation transcript:

1 MaxNet and TCP Reno/RED on mice traffic Khoa Truong Phan Ho Chi Minh city University of Technology (HCMUT)

Faculty of Computer Science and Engineering – HCMUT 2 Outline  Introduction  Overview of TCP Congestion Control  TCP Reno/RED and MaxNet TCP  Experiment and Evaluation

Faculty of Computer Science and Engineering – HCMUT 3 Introduction Figure 1. Traffic jam → Traffic on the Internet will be like this if we don’t have an efficient mechanism to avoid congestion.

Faculty of Computer Science and Engineering – HCMUT 4 Overview of TCP Congestion Control Congestion collapse TCP Vegas TCP Tahoe TCP Reno TCP NewReno FAST TCP 2006 Maxnet TCP Figure 2. History of TCP Congestion Control Algorithms

Faculty of Computer Science and Engineering – HCMUT 5 ACK packets` Data packets TCP receiver TCP sender PP p1p1 p2p2 P’’ PP’ Router Source compute rate to transmit Link mark/drop packets P’’=P’+ P 1 P=P’’+ P 2 P’ P’’ PP P Rate P= ∑P i Sink sends ACK to source Overview of TCP Congestion Control (cont) Figure 3. TCP Congestion Control model

Faculty of Computer Science and Engineering – HCMUT 6 TCP Reno Figure 4. Demand function of TCP Reno AIMD (Additive Increase Multiplicative Decrease) mechanism:

Faculty of Computer Science and Engineering – HCMUT 7 TCP Reno (cont) Figure 5. Operation mode of TCP Reno

Faculty of Computer Science and Engineering – HCMUT 8 RED router (1)(2) (3) b min b max 2b ma x RED router defines two thresholds in the buffer: b min and b max. The probability of marking/dropping (p) as follows: Figure 6. Operation mode of RED router

Faculty of Computer Science and Engineering – HCMUT 9 TCP Reno/RED  Assuming that sending rate at source is and router drops the packets at the probability of. Every drop packet causes a negative ACK.  Based on AIMD, source increases window size by 1/w for each positive ACK and decreases window size by half for each negative ACK.  At equilibrium, window size adjustment equal to zero  From the dropping scheme of RED, each source always have backlog at least at one router.  Window size adjustment:

Faculty of Computer Science and Engineering – HCMUT 10 TCP Reno/RED (cont) Elephant traffic Mice traffic RED router Figure 7. Queuing delay of RED router

Faculty of Computer Science and Engineering – HCMUT 11 MaxNet TCP Source: Router: Figure 8. Operation mode of TCP Reno Demand function: µ < 100%

Faculty of Computer Science and Engineering – HCMUT 12 MaxNet TCP (cont) Figure 9. Operation mode of MaxNet TCP MaxStart Figure 10. Queuing delay of RED router

Faculty of Computer Science and Engineering – HCMUT 13 Experiment Figure 11. Experiment test bed  Pentium IV PCs (CPU 1.8GHz, 512MB RAM) are used  Dummynet router is used to configured end-to-end delay at 20ms

Faculty of Computer Science and Engineering – HCMUT 14 Experiment (cont)  Monitoring the queue at MaxNet router and Reno/RED router Figure 13(a). Queue at MaxNet router Figure 13(b). Queue at Reno/RED router

Faculty of Computer Science and Engineering – HCMUT 15 Experiment (cont)  Response time of HTTP connections in MaxNet and Reno/RED Figure 14. Response time of HTTP in Reno/RED vs. MaxNet

Faculty of Computer Science and Engineering – HCMUT 16 Experiment (cont) Figure 12(a). 1 FTP and 50 HTTP connections Figure 12(a). 1 FTP and 100 HTTP connections  Impact of new HTTP connections on throughput of elephant traffic Reno/RED and MaxNet

Faculty of Computer Science and Engineering – HCMUT 17 Conclusions  MaxNet clears buffer while Reno/RED always keeps a backlog in routers.  MaxNet has shorter response time for mice traffic than Reno/RED.  Arrival mice flows cause packet loss which degrades the throughput of elephant traffic.  MaxStart mechanism of MaxNet, using multi-bit signaling, controls mice flows to the target rate more quickly than TCP Reno.  For using MaxNet, source hosts, intermediate routers and destination hosts need to be upgraded.

Faculty of Computer Science and Engineering – HCMUT 18 References  MaxNet homepage:  Duc Nguyen, Jidong Wang, Lachlan L. H. Andrew and Sammy Chan, “MaxNet: A More Efficient Max-min Fair Allocation Scheme”, in Proc. Intl. Teletraffic Congress-19, Beijing China,  Bartek Wydrowski, Lachlan L.H. Andrew, Moshe Zukerman, "MaxNet: A Congestion Control Architecture for Scalable Networks",IEEE Communications Letters, vol. 7, no. 10, Oct. 2003, pp  Bartek Wydrowski, Lachlan L.H. Andrew, Iven M. Y. Mareels, "MaxNet: Faster Flow Control Convergence", NETWORKING 2004:

Faculty of Computer Science and Engineering – HCMUT 19 THANK YOU!

Faculty of Computer Science and Engineering – HCMUT 20 Demand function of MaxNet →  Achieve Max-min fairness  Stability

Faculty of Computer Science and Engineering – HCMUT 21 MaxMin Fairness MaxMin fairness allocation

Faculty of Computer Science and Engineering – HCMUT 22 Fairness bandwidth of MaxNet vs. Reno MaxNet TCP Reno

Faculty of Computer Science and Engineering – HCMUT 23 Elephant traffic MaxNet vs. TCP Reno MaxNet Reno

Faculty of Computer Science and Engineering – HCMUT 24 Throughput of MaxNet Throughput of Reno