2  Something “feels the same” regardless of scale 4 What is that???

Slides:



Advertisements
Similar presentations
Random Processes Introduction (2)
Advertisements

Internet Measurement Conference 2003 Source-Level IP Packet Bursts: Causes and Effects Hao Jiang Constantinos Dovrolis (hjiang,
Estimation and identification of long-range dependence in Internet traffic Thomas Karagiannis University of California,
Doc.: IEEE /1216r1 Submission November 2009 BroadcomSlide 1 Internet Traffic Modeling Date: Authors: NameAffiliationsAddressPhone .
Copyright © 2005 Department of Computer Science CPSC 641 Winter Self-Similar Network Traffic The original paper on network traffic self-similarity.
Lesson 17: Models for Traffic Sources Giovanni Giambene Queuing Theory and Telecommunications: Networks and Applications 2nd edition, Springer All rights.
Network and Service Assurance Laboratory Analysis of self-similar Traffic Using Multiplexer & Demultiplexer Loaded with Heterogeneous ON/OFF Sources Huai.
2014 Examples of Traffic. Video Video Traffic (High Definition) –30 frames per second –Frame format: 1920x1080 pixels –24 bits per pixel  Required rate:
Finding Self-similarity in People Opportunistic Networks Ling-Jyh Chen, Yung-Chih Chen, Paruvelli Sreedevi, Kuan-Ta Chen Chen-Hung Yu, Hao Chu.
STAT 497 APPLIED TIME SERIES ANALYSIS
Chen Chu South China University of Technology. 1. Self-Similar process and Multi-fractal process There are 3 different definitions for self-similar process.
1 Alberto Montanari University of Bologna Simulation of synthetic series through stochastic processes.
1 Self-Similar Ethernet LAN Traffic Carey Williamson University of Calgary.
CMPT 855Module Network Traffic Self-Similarity Carey Williamson Department of Computer Science University of Saskatchewan.
On the Self-Similar Nature of Ethernet Traffic - Leland, et. Al Presented by Sumitra Ganesh.
1 17. Long Term Trends and Hurst Phenomena From ancient times the Nile river region has been known for its peculiar long-term behavior: long periods of.
October 14, 2002MASCOTS Workload Characterization in Web Caching Hierarchies Guangwei Bai Carey Williamson Department of Computer Science University.
Finding Self-similarity in Opportunistic People Networks Yung-Chih Chen 1 Ling-Jyh Chen 1, Yung-Chih Chen 1, Tony Sun 2 Paruvelli Sreedevi 1, Kuan-Ta Chen.
Statistics & Modeling By Yan Gao. Terms of measured data Terms used in describing data –For example: “mean of a dataset” –An objectively measurable quantity.
A gentle introduction to fluid and diffusion limits for queues Presented by: Varun Gupta April 12, 2006.
Small scale analysis of data traffic models B. D’Auria - Eurandom joint work with S. Resnick - Cornell University.
Probability By Zhichun Li.
A Nonstationary Poisson View of Internet Traffic T. Karagiannis, M. Molle, M. Faloutsos University of California, Riverside A. Broido University of California,
Self-Similarity in Network Traffic Kevin Henkener 5/29/2002.
1 Interesting Links
1 Using A Multiscale Approach to Characterize Workload Dynamics Characterize Workload Dynamics Tao Li June 4, 2005 Dept. of Electrical.
Economics 20 - Prof. Anderson
Origins of Long Range Dependence Myths and Legends Aleksandar Kuzmanovic 01/08/2001.
3-1 Introduction Experiment Random Random experiment.
Long Range Dependent Traffic and Leaky Buckets CS215-winter ’01 Demetrios Laios 3/22/2001.
Self-Similar through High-Variability: Statistical Analysis of Ethernet LAN Traffic at the Source Level Walter Willinger, Murad S. Taqqu, Robert Sherman,
CS 6401 Network Traffic Characteristics Outline Motivation Self-similarity Ethernet traffic WAN traffic Web traffic.
Internet Traffic Modeling Poisson Model vs. Self-Similar Model By Srividhya Chandrasekaran Dept of CS University of Houston.
Self-Similar Traffic COMP5416 Advanced Network Technologies.
Self-Similarity of Network Traffic Presented by Wei Lu Supervised by Niclas Meier 05/
Traffic modeling and Prediction ----Linear Models
1 Chapters 9 Self-SimilarTraffic. Chapter 9 – Self-Similar Traffic 2 Introduction- Motivation Validity of the queuing models we have studied depends on.
Panel Topic: After Long Range Dependency (LRD) discoveries, what are the lessons learned so far to provide QoS for Internet advanced applications David.
References for M/G/1 Input Process
Network Traffic Modeling Punit Shah CSE581 Internet Technologies OGI, OHSU 2002, March 6.
Traffic Modeling.
Simulation Output Analysis
“A non parametric estimate of performance in queueing models with long-range correlation, with applications to telecommunication” Pier Luigi Conti, Università.
1 FARIMA(p,d,q) Model and Application n FARIMA Models -- fractional autoregressive integrated moving average n Generating FARIMA Processes n Traffic Modeling.
1 Statistical Distribution Fitting Dr. Jason Merrick.
0 K. Salah 2. Review of Probability and Statistics Refs: Law & Kelton, Chapter 4.
Link Dimensioning for Fractional Brownian Input Chen Jiongze PhD student, Electronic Engineering Department, City University of Hong Kong Supported by.
1 Self Similar Traffic. 2 Self Similarity The idea is that something looks the same when viewed from different degrees of “magnification” or different.
Week 21 Stochastic Process - Introduction Stochastic processes are processes that proceed randomly in time. Rather than consider fixed random variables.
A Nonstationary Poisson View of Internet Traffic Thomas Karagiannis joint work with Mart Molle, Michalis Faloutsos, Andre Broido.
K. Ensor, STAT Spring 2004 Memory characterization of a process How would the ACF behave for a process with no memory? What is a short memory series?
MODELING THE SELF-SIMILAR BEHAVIOR OF PACKETIZED MPEG-4 VIDEO USING WAVELET-BASED METHODS Dogu Arifler and Brian L. Evans The University of Texas at Austin.
Performance Evaluation of Long Range Dependent Queues Performance Evaluation of Long Range Dependent Queues Chen Jiongze Supervisor: Moshe ZukermanCo-Supervisor:
1 EE571 PART 3 Random Processes Huseyin Bilgekul Eeng571 Probability and astochastic Processes Department of Electrical and Electronic Engineering Eastern.
Risk Analysis Workshop April 14, 2004 HT, LRD and MF in teletraffic1 Heavy tails, long memory and multifractals in teletraffic modelling István Maricza.
CHAPTER 2.3 PROBABILITY DISTRIBUTIONS. 2.3 GAUSSIAN OR NORMAL ERROR DISTRIBUTION  The Gaussian distribution is an approximation to the binomial distribution.
Notices of the AMS, September Internet traffic Standard Poisson models don’t capture long-range correlations. Poisson Measured “bursty” on all time.
1 Interesting Links. On the Self-Similar Nature of Ethernet Traffic Will E. Leland, Walter Willinger and Daniel V. Wilson BELLCORE Murad S. Taqqu BU Analysis.
Queuing Theory and Traffic Analysis
CMPE 252A: Computer Networks
Internet Traffic Modeling
Interesting Links.
Minimal Envelopes.
Notices of the AMS, September 1998
Self-similar Distributions
Mark E. Crovella and Azer Bestavros Computer Science Dept,
STOCHASTIC HYDROLOGY Random Processes
Presented by Chun Zhang 2/14/2003
CHAPTER- 3.1 ERROR ANALYSIS.
CPSC 641: Network Traffic Self-Similarity
Presentation transcript:

2

 Something “feels the same” regardless of scale 4 What is that???

 Something “feels the same” regardless of scale 5 Self-similar in nature

 Something “feels the same” regardless of scale 6 The Koch snowflake fractal

 Something “feels the same” regardless of scale 7 The Koch snowflake fractal

 Something “feels the same” regardless of scale 8 The Koch snowflake fractal

 Something “feels the same” regardless of scale 9

10 Categories:  Exact self-similarity: Strongest Type  Approximate self-similarity: Loose Form  Statistical self-similarity: Weakest Type

11 Approximate self-similarity: Recognisably similar but not exactly so. e.g. Mandelbrot set Statistical self-similarity: Only numerical or statistical measures that are preserved across scales

In case of Stochastic Objects e.g. time-series Self-similarity is used in the distributional sense 12

 Recently, network packet traffic has been identified as being self-similar.  Current network traffic modeling using Poisson distributing (etc.) does not take into account the self-similar nature of traffic.  This leads to inaccurate modeling of network traffic. 13

 A Poisson process  When observed on a fine time scale will appear bursty  When aggregated on a coarse time scale will flatten (smooth) to white noise  A Self-Similar (fractal) process  When aggregated over wide range of time scales will maintain its bursty characteristic 14

15 packets per time unit Ethernet traffic August’89 trace

16

17

18

19 Bursty Data Streams Aggregation Smooth Pattern Streams Bursty Data Streams Aggregation Bursty Aggregate Streams Reality (self-similar): Current Model: Consequence: Inaccuracy

 Long-range Dependence  autocorrelation decays slowly  Hurst Parameter  Developed by Harold Hurst (1965)  H is a measure of “burstiness” ▪ also considered a measure of self-similarity  0 < H < 1  H increases as traffic increases ▪ i.e., traffic becomes more self-similar 20

 X = (X t : t = 0, 1, 2, ….) is covariance stationary random process (i.e. Cov(X t,X t+k ) does not depend on t for all k)  Let X (m) ={X k (m) } denote the new process obtained by averaging the original series X in non-overlapping sub-blocks of size m.  Mean , variance  2  Suppose that Autocorrelation Function r(k)  k -β, 0<β<1 21 e.g. X(1)= 4,12,34,2,-6,18,21,35 Then X(2)=8,18,6,28 X(4)=13,17

 X is exactly second-order self-similar if  The aggregated processes have the same autocorrelation structure as X. i.e.  r (m) (k) = r(k), k  0 for all m =1,2, …  X is asymptotically second-order self-similar if the above holds when [ r (m) (k)  r(k), m    Most striking feature of self-similarity: Correlation structures of the aggregated process do not degenerate as m   22

23 lag ACF

24

 Correlation structures of their aggregated processes degenerate as m    i.e. r (m) (k)  0 as m  for k = 1,2,3,...  Short Range Dependence Processes:  Exponential Decay of autocorrelations  i.e. r(k) ~ p k, as k  , 0 < p < 1  Summation is finite 25

 Processes with Long Range Dependence are characterized by an autocorrelation function that decays hyperbolically as k increases  Important Property: This is also called non-summability of correlation 26

 The intuition behind long-range dependence:  While high-lag correlations are all individually small, their cumulative affect is important  Gives rise to features drastically different from conventional short-range dependent processes 27

 Hurst Parameter H, 0.5 < H < 1  Three approaches to estimate H (Based on properties of self-similar processes)  Variance Analysis of aggregated processes  Rescaled Range (R/S) Analysis for different block sizes: time domain analysis  Periodogram Analysis: frequency domain analysis (Whittle Estimator) 28

 Variance of aggregated processes decays as:  Var(X (m) ) = am -b as m  infinite,  For short range dependent processes (e.g. Poisson Process):  Var(X (m) ) = am -1 as m  infinite,  Plot Var(X (m) ) against m on a log-log plot  Slope > -1 indicative of self-similarity 29

30 Slope=-1 Slope=-0.7

31 where For a given set of observations, Rescaled Adjusted Range or R/S statistic is given by

 X k = 14,1,3,5,10,3  Mean = 36/6 = 6  W 1 =14-(1*6 )=8  W 2 =15-(2*6 )=3  W 3 =18-(3*6 )=0  W 4 =23-(4*6 )=-1  W 5 =33-(5*6 )=3  W 6 =36-(6*6 )=0 32 R/S = 1/S*[8-(-1)] = 9/S

 For self-similar data, rescaled range or R/S statistic grows according to cn H  H = Hurst Paramater, > 0.5  For short-range processes,  R/S statistic ~ dn 0.5  History: The Nile river  In the ’s, Harold Edwin Hurst studied the 800-year record of flooding along the Nile river.  (yearly minimum water level)  Finds long-range dependence. 33

34 Slope = 1.0 Slope = 0.5 Slope = 0.79

 Provides a confidence interval  Property: Any long range dependent process approaches fractional Gaussian noise (FGN), when aggregated to a certain level  Test the aggregated observations to ensure that it has converged to the normal distribution 35

 Self-similarity manifests itself in several equivalent fashions:  Non-degenerate autocorrelations  Slowly decaying variance  Long range dependence  Hurst effect 36

 Leland and Wilson collected hundreds of millions of Ethernet packets without loss and with recorded time-stamps accurate to within 100µs.  Data collected from several Ethernet LAN’s at the Bellcore Morristown Research and Engineering Center at different times over the course of approximately 4 years. 38

39

40 H=0.5 H=1 Estimate H  0.8

41 Higher Traffic, Higher H High Traffic Mid Traffic Low Traffic 1.3%-10.4% 3.4%-18.4% 5.0%-30.7% Packets

 Observation shows “contrary to Poisson”  Network UtilizationH 42 As number of Ethernet users increases, the resulting aggregate traffic becomes burstier instead of smoother

 Pre-1990: host-to-host workgroup traffic  Post-1990: Router-to-router traffic  Low period router-to-router traffic consists mostly of machine-generated packets  Tend to form a smoother arrival stream, than low period host-to-host traffic 43

 Ethernet LAN traffic is statistically self-similar  H : the degree of self-similarity  H : a function of utilization  H : a measure of “burstiness”  Models like Poisson are not able to capture self-similarity 44

46

 The superposition of many ON/OFF sources whose ON-periods and OFF-periods exhibit the Noah Effect produces aggregate network traffic that features the Joseph Effect. 47 Also known as packet train models Noah Effect: high variability or infinite variance Joseph Effect: Self-similar or long-range dependent traffic

 Traditional traffic models: finite variance ON/OFF source models  Superposition of such sources behaves like white noise, with only short range correlations 48

 Questions related to self-similarity can be reduced to practical implications of Noah Effect  Queuing and Network performance  Network Congestion Controls  Protocol Analysis 49

 The Queue Length distribution  Traditional (Markovian) traffic: decreases exponentially fast  Self-similar traffic: decreases much more slowly  Not accounting for Joseph Effect can lead to overly optimistic performance 50 Effect of H (Burstiness)

 How to design the buffer size?  Trade-off between Packet Lose and Packet Delay 51

52 Packet LosePacket Delay Short Range DependenceDecrease ExponentiallyFixed Limit Long Range DependenceDecrease SlowlyAlways Increase Compare SRD and LRD when increase buffer size

 Protocol design should take into account knowledge about network traffic such as the presence or absence of the self-similarity. 53  Parsimonious Models Small number of parameters Every parameter has a physically meaningful interpretation e.g. Mean , Variance  2, H Doesn’t quantify the effects of various factors in traffic

 Demonstrated the existence of self-similarity in Ethernet Traffic irrespective of time scales  Proposed the degree of self-similarity can be measured by Hurst parameter H (higher H implies burstier traffic)  Illustrated the difference between the self-similar and standard models  Explained Importance of self similarity in design, control, performance analysis 54

55