Multiscale Traffic Processing Techniques for Network Inference and Control R. Baraniuk R. Nowak E. Knightly R. Riedi V. Ribeiro S. Sarvotham A. Keshavarz.

Slides:



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

Internet Measurement Conference 2003 Source-Level IP Packet Bursts: Causes and Effects Hao Jiang Constantinos Dovrolis (hjiang,
Edge-based Network Modeling and Inference
Collaborators: Mark Coates, Rui Castro, Ryan King, Mike Rabbat, Yolanda Tsang, Vinay Ribeiro, Shri Sarvotham, Rolf Reidi Network Bandwidth Estimation and.
Connection-level Analysis and Modeling of Network Traffic understanding the cause of bursts control and improve performance detect changes of network state.
Available Bandwidth Estimation Manish Jain Networking and Telecom Group CoC, Georgia Tech.
pathChirp Efficient Available Bandwidth Estimation
Ningning HuCarnegie Mellon University1 Optimizing Network Performance In Replicated Hosting Peter Steenkiste (CMU) with Ningning Hu (CMU), Oliver Spatscheck.
Aleksandar Kuzmanovic and Edward W. Knightly Rice Networks Group Measuring Service in Multi-Class Networks.
Path Optimization in Computer Networks Roman Ciloci.
End-to-End Available Bandwidth: Measurement Methodology, Dynamics, and Relation with TCP Throughput Manish Jain Constantinos Dovrolis SIGCOMM 2002 Presented.
XCP: Congestion Control for High Bandwidth-Delay Product Network Dina Katabi, Mark Handley and Charlie Rohrs Presented by Ao-Jan Su.
One More Bit Is Enough Yong Xia, RPI Lakshminarayanan Subramanian, UCB Ion Stoica, UCB Shivkumar Kalyanaraman, RPI SIGCOMM’05, August 22-26, 2005, Philadelphia,
Farsighted Congestion Controllers Milan Vojnović Microsoft Research Cambridge, United Kingdom Collaborators: Dinan Gunawardena (MSRC), Peter Key (MSRC),
Network Traffic Measurement and Modeling CSCI 780, Fall 2005.
A Nonstationary Poisson View of Internet Traffic T. Karagiannis, M. Molle, M. Faloutsos University of California, Riverside A. Broido University of California,
WBest: a Bandwidth Estimation Tool for IEEE Wireless Networks Presented by Feng Li Mingzhe Li, Mark Claypool, and.
1 TCP-LP: A Distributed Algorithm for Low Priority Data Transfer Aleksandar Kuzmanovic, Edward W. Knightly Department of Electrical and Computer Engineering.
1 End-to-End Detection of Shared Bottlenecks Sridhar Machiraju and Weidong Cui Sahara Winter Retreat 2003.
1 Emulating AQM from End Hosts Presenters: Syed Zaidi Ivor Rodrigues.
1 Active Probing for Available Bandwidth Estimation Sridhar Machiraju UC Berkeley OASIS Retreat, Jan 2005 Joint work with D.Veitch, F.Baccelli, A.Nucci,
Rethinking Internet Traffic Management: From Multiple Decompositions to a Practical Protocol Jiayue He Princeton University Joint work with Martin Suchara,
Origins of Long Range Dependence Myths and Legends Aleksandar Kuzmanovic 01/08/2001.
TCP Congestion Control
Low-Rate TCP-Targeted Denial of Service Attacks Presenter: Juncao Li Authors: Aleksandar Kuzmanovic Edward W. Knightly.
Inline Path Characteristic Estimation to Improve TCP Performance in High Bandwidth-Delay Networks HIDEyuki Shimonishi Takayuki Hama Tutomu Murase Cesar.
Bandwidth Estimation: Metrics Mesurement Techniques and Tools By Ravi Prasad, Constantinos Dovrolis, Margaret Murray and Kc Claffy IEEE Network, Nov/Dec.
Self-Similarity of Network Traffic Presented by Wei Lu Supervised by Niclas Meier 05/
1 Chapters 9 Self-SimilarTraffic. Chapter 9 – Self-Similar Traffic 2 Introduction- Motivation Validity of the queuing models we have studied depends on.
Alok Shriram and Jasleen Kaur Presented by Moonyoung Chung Empirical Evaluation of Techniques for Measuring Available Bandwidth.
Multiscale Traffic Processing Techniques for Network Inference and Control R. Baraniuk R. Nowak E. Knightly R. Riedi V. Ribeiro S. Sarvotham A. Keshavarz.
Multiscale Traffic Processing Techniques for Network Inference and Control Richard Baraniuk Edward Knightly Robert Nowak Rolf Riedi Rice University INCITE.
Estimating Bandwidth of Mobile Users Sept 2003 Rohit Kapoor CSD, UCLA.
Development of a New Efficient and Accurate Available Bandwidth Estimation Method Péter Hága Attila Pásztor István Csabai Darryl Veitch Viktória Hunyadi.
Comparison of Public End-to-End Bandwidth Estimation tools on High-Speed Links Alok Shriram, Margaret Murray, Young Hyun, Nevil Brownlee, Andre Broido,
Comparison of Public End-to-End Bandwidth Estimation tools on High- Speed Links Alok Shriram, Margaret Murray, Young Hyun, Nevil Brownlee, Andre Broido,
DoE SciDAC high-performance networking research project: INCITE INCITE.rice.edu 2004 Technical Challenges INCITE R. Baraniuk, E. Knightly, R. Nowak, R.
Congestion Control in CSMA-Based Networks with Inconsistent Channel State V. Gambiroza and E. Knightly Rice Networks Group
Requirements for Simulation and Modeling Tools Sally Floyd NSF Workshop August 2005.
Opportunistic Traffic Scheduling Over Multiple Network Path Coskun Cetinkaya and Edward Knightly.
Multiplicative Wavelet Traffic Model and pathChirp: Efficient Available Bandwidth Estimation Vinay Ribeiro.
Optimal Sampling Strategies for Multiscale Stochastic Processes Vinay Ribeiro Rolf Riedi, Rich Baraniuk (Rice University)
PathChirp Spatio-Temporal Available Bandwidth Estimation Vinay Ribeiro Rolf Riedi, Richard Baraniuk Rice University.
Spatio-Temporal Available Bandwidth Estimation Vinay Ribeiro Rolf Riedi, Richard Baraniuk Rice University.
Rice University, SPiN Group spin.rice.edu Multiscale Traffic Processing Techniques for Network Inference and Control R. Baraniuk R. Nowak E. Knightly R.
Analysis of RED Goal: impact of RED on loss and delay of bursty (TCP) and less bursty or smooth (UDP) traffic RED eliminates loss bias against bursty traffic.
Rice University, SPiN Group spin.rice.edu Multiscale Traffic Processing Techniques for Network Inference and Control R. Baraniuk R. Nowak E. Knightly R.
The Macroscopic behavior of the TCP Congestion Avoidance Algorithm.
DoE SciDAC high-performance networking research project: INCITE INCITE.rice.edu 2004 Technical Challenges INCITE R. Baraniuk, E. Knightly, R. Nowak, R.
INCITE: Traffic Processing Techniques for Network Inference and Control Effort 1: Chirp Probing Objective : Reduced complexity, multiscale link models.
Péter Hága Eötvös Loránd University, Hungary European Conference on Complex Systems 2008 Jerusalem, Israel.
Development of a QoE Model Himadeepa Karlapudi 03/07/03.
Grid Computing slide to be used anywhere Harness global resources to improve performance.
Multiscale Tools for Networking Vinay J. Ribeiro Ph.D. Thesis Proposal.
Internet Measurement and Analysis Vinay Ribeiro Shriram Sarvotham Rolf Riedi Richard Baraniuk Rice University.
PathChirp & STAB Measuring Available Bandwidth and Locating Bottlenecks in Packet Networks Vinay Ribeiro Rolf Riedi, Richard Baraniuk Rice University spin.rice.edu.
1 Transport Layer: Basics Outline Intro to transport UDP Congestion control basics.
PathChirp Efficient Available Bandwidth Estimation Vinay Ribeiro Rice University Rolf Riedi Rich Baraniuk.
Notices of the AMS, September Internet traffic Standard Poisson models don’t capture long-range correlations. Poisson Measured “bursty” on all time.
Bandwidth estimation: metrics, measurement techniques, and tools Presenter: Yuhang Wang.
Queuing Analysis of Tree-Based LRD Traffic Models Vinay J. Ribeiro R. Riedi, M. Crouse, R. Baraniuk.
Bandwidth Estimation of a Network Path ET-4285 Measuring & Simulating the internet Bandwidth Estimation of a Network Path Group 4: S. Ngabonziza Rugemintwaza.
Bandwidth Estimation: Metrics Measurement Techniques and Tools
Monitoring Persistently Congested Internet Links
Topics discussed in this section:
Measuring Service in Multi-Class Networks
Amogh Dhamdhere, Hao Jiang and Constantinos Dovrolis
TCP Congestion Control
pathChirp Efficient Available Bandwidth Estimation
pathChirp Efficient Available Bandwidth Estimation
Presentation transcript:

Multiscale Traffic Processing Techniques for Network Inference and Control R. Baraniuk R. Nowak E. Knightly R. Riedi V. Ribeiro S. Sarvotham A. Keshavarz NMS PI meeting San Diego May 2003 SPiN.Rice.edu: Signal Processing in Networking

Rice University, SPiN Group spin.rice.edu Chirp Probing always-on, non-intrusive bandwidth estimation on-line decision anomaly detection Effort 1

Rice University, SPiN Group spin.rice.edu Milestones pathChirp May 2003 milestone: C code testing + distribution [completed] –Allen McIntosh (Telcordia): testing, useful comments, –SPAWAR (hosted Phuong Nguyen at Rice), –CAIDA (kc Claffy, Margaret Murray), –RPI (Shivkumar). –Demo: on-line estimation of bandwidth Towards Nov 2003 milestones: Integration of pathChirp –GaTech pdns (Riley/Fujimoto) [completed] –UIUC, JavaSim (Hou) [in progress] Validation in controllable environment [enabled, to be done] Towards May 2004 milestones: [basis provided, work to be done] Final tool and theory for bursty traffic over multiple hops

Rice University, SPiN Group spin.rice.edu Efficient probing: PathChirp Traditional probing paradigm: –Produce (light) congestion –PacketPair: Sample the traffic –Pathload: flood at variable rate intolerable level of congestion –TOPP: PacketPairs at variable spacing New: –PathChirp: Variable rate within a train of probes More efficient, light

Rice University, SPiN Group spin.rice.edu pathChirp Developments pathChirp …a real world tool …with improved performance –Increased queuing delay correlates with cross traffic on network path –Last excursion in chirp  link capacity –Weighted averaged of onset of excursions  available link resources Departure pattern Queuing against departure Methodology

Rice University, SPiN Group spin.rice.edu pathChirp Developments pathChirp …a real world tool …with improved performance –Queuing delay  cross traffic –Final excursion  link capacity –Averaged excursions  available resources …converges in a handful of RTTs Departure pattern Queuing against departure Methodology Number of chirps  12 chirps Real world experiments Estimation against true x-traffic Internet experiment

Rice University, SPiN Group spin.rice.edu PathChirp Performance pathChirp performs comparably to –PathLoad –PacketPair –TOPP …at smaller probing rate …more robust to bursty traffic Best paper at PAM2003 Ongoing work: –Exploit dispersion information captured in excursions to become robust against multiple hops pathChirp vs TOPP square error PathLoad converged after 6.7 Mb.5 Mb1 Mb

Rice University, SPiN Group spin.rice.edu Connection-level Analysis and Modeling of Network Traffic understanding the cause of bursts control and improve performance detect changes of network state Effort 2

Rice University, SPiN Group spin.rice.edu Milestones Alpha-Beta May 2003 milestone [completed]: C++ version of decomposition and analysis module Towards Nov 2003 milestone: Verification of alpha-beta hypothesis in wider range of topologies, protocols, applications [Analysis module ready; collection and analysis to be done] Collaborations with Telcordia and SLAC [initiated] Collaborations with CAIDA [pursuing] December 2004 milestones [to be done]: –Integration into simulators, verification in large simulation –Applications: alpha-bottleneck aware AQM, Admission control

Rice University, SPiN Group spin.rice.edu Non-Gaussianity and Dominance Connection level separation: –remove packets of the ONE strongest connection –Leaves “Gaussian” residual traffic Traffic components: –Alpha connections: high rate (> ½ bandwidth) –Beta connections: all the rest Overall traffic Residual traffic1 Strongest connection = + Mean 99%

Rice University, SPiN Group spin.rice.edu Simple Connection Taxonomy Bursts arise from large transfers over fast links.

Rice University, SPiN Group spin.rice.edu CWND or RTT? Correlation coefficient=0.68 Short RTT correlates directly with high rate and bursts. peak-rate (Bps) 1/RTT (1/s) Correlation coefficient= peak-rate (Bps) cwnd (B) Colorado State University trace, 300,000 packets BetaAlphaBetaAlpha

Rice University, SPiN Group spin.rice.edu Impact: Performance Beta Traffic rules the small Queues Alpha Traffic causes the large Queue-sizes (despite small Window Size) Alpha connections Queue-size overlapped with Alpha Peaks Total traffic

Rice University, SPiN Group spin.rice.edu Two models for alpha traffic Impact of alpha burst in two scenarios: Flow control at end hosts – TCP advertised window Congestion control at router – TCP congestion window

Modeling Alpha Traffic ON/OFF model revisited: High variability in connection rates (RTTs) Low rate = betaHigh rate = alpha fractional Gaussian noise stable Levy noise + = + + =

Rice University, SPiN Group spin.rice.edu Alpha-Beta Model of Traffic Model assumptions: –Total traffic = Alpha component + Beta component –Alpha and Beta are independent –Beta=fractional Brownian motion Alpha traffic: two scenarios –Flow control through thin or busy end-hosts ON-OFF Burst model –Congestion control allowing large CWND Self-Similar Burst model Methods of analysis –Self-similar traffic –Queue De-multiplexing –Variable service rate

Self-similar Burst Model Alpha component = self-similar stable –(limit of a few ON-OFF sources in the limit of fast time) This models heavy-tailed bursts –(heavy tailed files) TCP control: alpha CWND arbitrarily large –(short RTT, future TCP mutants) Analysis via De-Multiplexing: –Optimal setup of two individual Queues to come closest to aggregate Queue De-Multiplexing: Equal critical time-scales Q-tail Pareto Due to Levy noise Beta (top) + Alpha

Rice University, SPiN Group spin.rice.edu ON-OFF Burst Model Alpha traffic = High rate ON-OFF source (truncated) This models bi-modal bandwidth distribution TCP: bottleneck is at the receiver (flow control through advertised window) Current state of measured traffic Analysis: de-multiplexing and variable rate queue Beta (top) + Alpha Variable Service Rate Queue-tail Weibull (unaffected) unless rate of alpha traffic larger than capacity – average beta arrival and duration of alpha ON period heavy tailed

Rice University, SPiN Group spin.rice.edu Alpha traffic: Influence of TCP All Alpha connections show –Unusually small advertised window –Drastic drop in advertised window (sometimes to zero) –…which correlates with burst arrival  Flow controlled, Weibull Q-tails

Rice University, SPiN Group spin.rice.edu Separation on Connection Level Alpha connections: dominant. Properties: –Definition: Peak rate > mean arrival rate + 1 std dev –Few, light load –Responsible for violent bursts, large queuing delays –Typically short RTT –Typically FLOW-CONTROLLED (limited at receiver) Beta connections: Residual traffic –Main load –Gaussian, LRD –Typically limited at bottleneck link Future of empowered hosts and transfer protocols: –Higher peaks, larger bursts, longer queues

Rice University, SPiN Group spin.rice.edu Future work : Network/user-driven traffic model –Correlations between network and user –Through simulation and measurements assess impact of protocols, applications, clientele, end-host server –Performance parameters from network and user specifications pathChirp –Model based estimation meeting challenges of bursty traffic –Through simulation validate realism (multihop, bursty traffic) –Anomaly detection through chirp-web Current Collaborations & Tech Transfer –IP-tunneling, coordinated measurements (Telcordia) –Integration of PathChirp into network simulators (GaTech, UIUC) –Ready for integration into SPAWAR –Demystify self-similarity ( UC Riverside)

Rice University, SPiN Group spin.rice.edu spin.rice.edu