Analysis and Comparison of TCP Reno and TCP Vegas Review

Slides:



Advertisements
Similar presentations
Congestion Control and Fairness Models Nick Feamster CS 4251 Computer Networking II Spring 2008.
Advertisements

TCP Variants.
On Individual and Aggregate TCP Performance Lili Qiu Yin Zhang Srinivasan Keshav Cornell University 7th International Conference on Network Protocols Toronto,
TCP Vegas LAWRENCE S. BRAKMO SEAN W. O’MALLEY LARRY L. PETERSON PRESENTED TCP VEGAS IN 1994 PRESENTED BY CHUNG TRAN.
TCP Vegas: New Techniques for Congestion Detection and Control.
CSIT560 Internet Infrastructure: Switches and Routers Active Queue Management Presented By: Gary Po, Henry Hui and Kenny Chong.
Restricted Slow-Start for TCP William Allcock 1,2, Sanjay Hegde 3 and Rajkumar Kettimuthu 1,2 1 Argonne National Laboratory 2 The University of Chicago.
Network Border Patrol Celio Albuquerque, Brett J. Vickers and Tatsuya Suda Jaideep Vaidya CS590F Fall 2000.
Advanced Computer Networking Congestion Control for High Bandwidth-Delay Product Environments (XCP Algorithm) 1.
Congestion Control An Overview -Jyothi Guntaka. Congestion  What is congestion ?  The aggregate demand for network resources exceeds the available capacity.
TCP Stability and Resource Allocation: Part II. Issues with TCP Round-trip bias Instability under large bandwidth-delay product Transient performance.
Networks: Congestion Control1 Congestion Control.
EE689 Lecture 5 Review of last lecture More on HPF RED.
TCP Congestion Control TCP sources change the sending rate by modifying the window size: Window = min {Advertised window, Congestion Window} In other words,
A Switch-Based Approach to Starvation in Data Centers Alex Shpiner Joint work with Isaac Keslassy Faculty of Electrical Engineering Faculty of Electrical.
1 Emulating AQM from End Hosts Presenters: Syed Zaidi Ivor Rodrigues.
TCP Vegas Brakmo & Peterson. No change in TCP spec, merely an alternative implementation –Changes needed only at sender side Main finding –Vegas achieves.
TCP Congestion Control
Congestion Control for High Bandwidth-Delay Product Environments Dina Katabi Mark Handley Charlie Rohrs.
Adaptive Control for TCP Flow Control Thesis Presentation Amir Maor.
TCP Vegas Kulan Kao 2006/3/25.
ACN: RED paper1 Random Early Detection Gateways for Congestion Avoidance Sally Floyd and Van Jacobson, IEEE Transactions on Networking, Vol.1, No. 4, (Aug.
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.
Queueing and Active Queue Management Aditya Akella 02/26/2007.
TCP with Variance Control for Multihop IEEE Wireless Networks Jiwei Chen, Mario Gerla, Yeng-zhong Lee.
What is TCP? Connection-oriented reliable transfer Stream paradigm
Hybrid Modeling of TCP Congestion Control João P. Hespanha, Stephan Bohacek, Katia Obraczka, Junsoo Lee University of Southern California.
Compound TCP in NS-3 Keith Craig 1. Worcester Polytechnic Institute What is Compound TCP? As internet speeds increased, the long ‘ramp’ time of TCP Reno.
TCP: Transmission Control Protocol Part II : Protocol Mechanisms Computer Network System Sirak Kaewjamnong Semester 1st, 2004.
1 SIGCOMM ’ 03 Low-Rate TCP-Targeted Denial of Service Attacks A. Kuzmanovic and E. W. Knightly Rice University Reviewed by Haoyu Song 9/25/2003.
CS640: Introduction to Computer Networks Aditya Akella Lecture 15 TCP – III Reliability and Implementation Issues.
1 Analysis of a window-based flow control mechanism based on TCP Vegas in heterogeneous network environment Hiroyuki Ohsaki Cybermedia Center, Osaka University,
H. OhsakiITCom A control theoretical analysis of a window-based flow control mechanism for TCP connections with different propagation delays Hiroyuki.
Chapter 11.4 END-TO-END ISSUES. Optical Internet Optical technology Protocol translates availability of gigabit bandwidth in user-perceived QoS.
© Janice Regan, CMPT 128, CMPT 371 Data Communications and Networking Congestion Control 0.
Congestion Avoidance Created by M Bateman, A Ruddle & C Allison As part of the TCP View project.
Increasing TCP's CWND based on Throughput draft-you-iccrg-throughput-based-cwnd-increasing-00 Jianjie You IETF92 Dallas.
Probabilistic Congestion Control for Non-Adaptable Flows Jörg Widmer, Martin Mauve, Jan Peter Damm (NOSSDAV’02) Presented by Ankur Upadhyaya for CPSC 538A.
TCP over Wireless PROF. MICHAEL TSAI 2016/6/3. TCP Congestion Control (TCP Tahoe) Only ACK correctly received packets Congestion Window Size: Maximum.
Other Methods of Dealing with Congestion
Advanced Computer Networks
Window Control Adjust transmission rate by changing Window Size
David Wetherall Spring 2000
TCP Vegas Congestion Control Algorithm
TCP Vegas: New Techniques for Congestion Detection and Avoidance
Topics discussed in this section:
TCP Vegas: New Techniques for Congestion Detection and Avoidance
Chapter 6 TCP Congestion Control
Congestion Control and
COMP 431 Internet Services & Protocols
Introduction to Congestion Control
CS 268: Lecture 6 Scott Shenker and Ion Stoica
TCP Vegas: New Techniques for Congestion Detection and Avoidance
Chapter 6 Congestion Avoidance
Generalizing The Network Performance Interference Problem
TCP Congestion Control
TCP-LP Distributed Algorithm for Low-Priority Data Transfer
Queue Dynamics with Window Flow Control
TCP Westwood(+) Protocol Implementation in ns-3
Lecture 19 – TCP Performance
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.
Other Methods of Dealing with Congestion
Other Methods of Dealing with Congestion
Chapter 6 TCP Congestion Control
COMP/ELEC 429/556 Introduction to Computer Networks
CS640: Introduction to Computer Networks
TCP: Transmission Control Protocol Part II : Protocol Mechanisms
Review of Internet Protocols Transport Layer
Lecture 6, Computer Networks (198:552)
Presentation transcript:

Analysis and Comparison of TCP Reno and TCP Vegas Review 10-14-04 Jack Meier Discussion leader

Discussion Main Idea – number of bytes in transmit is directly proportional to the expected throughput Well written but figures hard to read Points out both good and bad points Paper does not provide original work There are much better papers that come to the same conclusions Well known protocol and characteristics repeated- TCP Reno, that is currently used as a congestion avoidance mechanism, has some sufficient drawbacks, such as leaving little room for other connections, dependence on the connection delay in terms of fairness, loss causing that reduces throughput. Using TCP Vegas as an alternative, together with its improvements, can avoid several TCP Reno drawbacks. TCP is not a synchronized rate based control scheme The paper “TCP Vegas: New Techniques for Congestion Detection and Avoidance” has a more extensive analysis and identifies most of the key points more clearly Math could be cleaner and unnecessarily complicated Consensus of Students: Lower ½ of 13 papers – definitely not the best paper

Discussion for the Best Paper Well-written summary of the TCP Reno and Vegas algorithms Use RED routers to encourage the deployment of TCP Vegas. RED gateway drops packets thereby decreasing congestion mechanism to avoid persistent congestion in TCP Vegas. Proves that TCP Vegas can be made to work in the Internet and shows how Provides a mathematical model to show that TCP Vegas does not suffer from being "Delay" biased as TCP Reno. simulation results to show increasing link delays has little impact on BW given to a Source for TCP Vegas Increasing link delays for TCP Reno causes the Source associated with that link to get a decrease in BW that is almost linear with increases in link delay TCP Reno causes TCP Vegas to get less than its fair share of network BW Presents some of the draw backs of the TCP Vegas congestion control mechanism and possible first cut solutions Rerouting causing a link to have a increased associated delay. This is misinterprated by TCP Vegas to be a sign of congestion thereby incorrectly decreasing its window size sol: Sample N packets periodically to reset the baseRTT used On initial connection to a congested network TCP Vegas will interpret the RTT values as the baseRTT and push data into the network too quickly thereby making the congestion worse Analytically shows that TCP Vegas is much fairer than TCP Reno to connections with longer end-to-end delays Proves fairness of TCP Vegas with itself

Not Recommended to be the Best Paper A significant portion of the paper simply repeats results from earlier work Earlier papers are better and address nearly all the same points Rerouting and Persistent Congestion are briefly brought up; however, they are ignored in the remainder of the paper. TCP Vegas has major problems if wrong interpretation by the source is made due to topology changes that result in new minimum RTT propagation time changing Continuous congestion Dependence of TCP Vegas on RTT-accurate estimation Propagation delay of a route may change with topology No analysis of TCP Vegas for slow start

Criticisms A vast majority of the paper seems to only summarizing what was written in other papers, and very little appears to be providing any novel information It is ambiguous, its simulations superficial and its conclusion premature. It doesn't adequately explain several relevant points. The amount of buffer needed grows linearly with the number of flows, this causes a major scalability problem for TCP Vegas. Should the paper make this the main focus? The RED gateway is not a solution merely a fix to a known problem Does not explain the way you would implement TCP Vegas Mathematical definitions and proofs are weak

Questions Do you believe that the RED gateway legitimately solves the issues with making TCP Reno more compatible with TCP Vegas? How is the convergence point related to the queue size? Is the assumption that simplifies alpha and beta valid? (quasi-static) – discussed in presentation Is the assumption valid that queue size and throughput is static? How do you adjust the fairness line of the window sized pairs such that the user throughput is the same? Do you think implementing a DRR scheduler to instill fairness will work?