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.

Slides:



Advertisements
Similar presentations
Doc.: IEEE /1216r1 Submission November 2009 BroadcomSlide 1 Internet Traffic Modeling Date: Authors: NameAffiliationsAddressPhone .
Advertisements

Copyright © 2005 Department of Computer Science CPSC 641 Winter Self-Similar Network Traffic The original paper on network traffic self-similarity.
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:
Computer Science Generating Streaming Access Workload for Performance Evaluation Shudong Jin 3nd Year Ph.D. Student (Advisor: Azer Bestavros)
STAT 497 APPLIED TIME SERIES ANALYSIS
2  Something “feels the same” regardless of scale 4 What is that???
1 Network Traffic Measurement and Modeling Carey Williamson Department of Computer Science University of Calgary.
1 LAN Traffic Measurements Carey Williamson Department of Computer Science University of Calgary.
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.
Error Propagation. Uncertainty Uncertainty reflects the knowledge that a measured value is related to the mean. Probable error is the range from the mean.
Statistics & Modeling By Yan Gao. Terms of measured data Terms used in describing data –For example: “mean of a dataset” –An objectively measurable quantity.
On the Constancy of Internet Path Properties Yin Zhang, Nick Duffield AT&T Labs Vern Paxson, Scott Shenker ACIRI Internet Measurement Workshop 2001 Presented.
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.
無線區域網路中自我相似交通流量之 成因與效能評估 The origin and performance impact of self- similar traffic for wireless local area networks 報 告 者:林 文 祺 指導教授:柯 開 維 博士.
Network Traffic Measurement and Modeling CSCI 780, Fall 2005.
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
Copyright © 2005 Department of Computer Science CPSC 641 Winter LAN Traffic Measurements Some of the first network traffic measurement papers were.
Origins of Long Range Dependence Myths and Legends Aleksandar Kuzmanovic 01/08/2001.
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.
1 10. Joint Moments and Joint Characteristic Functions Following section 6, in this section we shall introduce various parameters to compactly represent.
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/
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
SIMULATION MODELING AND ANALYSIS WITH ARENA
Traffic Modeling.
1 FARIMA(p,d,q) Model and Application n FARIMA Models -- fractional autoregressive integrated moving average n Generating FARIMA Processes n Traffic Modeling.
0 K. Salah 2. Review of Probability and Statistics Refs: Law & Kelton, Chapter 4.
COMPSAC'14 - N. Larrieu /07/ How to generate realistic network traffic? Antoine VARET and Nicolas LARRIEU COMPSAC – Vasteras – July the 23.
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.
Review of Probability. Important Topics 1 Random Variables and Probability Distributions 2 Expected Values, Mean, and Variance 3 Two Random Variables.
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.
1 Hailuoto Workshop A Statistician ’ s Adventures in Internetland J. S. Marron Department of Statistics and Operations Research University of North Carolina.
1 Internet Traffic Measurement and Modeling Carey Williamson Department of Computer Science University of Calgary.
Notices of the AMS, September Internet traffic Standard Poisson models don’t capture long-range correlations. Poisson Measured “bursty” on all time.
Ó 1998 Menascé & Almeida. All Rights Reserved.1 Part VIII Web Performance Modeling (Book, Chapter 10)
Queuing Theory and Traffic Analysis
CMPE 252A: Computer Networks
Internet Traffic Modeling
Interesting Links.
Minimal Envelopes.
CPSC 641: LAN Measurement Carey Williamson
Notices of the AMS, September 1998
Self-similar Distributions
Mark E. Crovella and Azer Bestavros Computer Science Dept,
Presented by Chun Zhang 2/14/2003
Carey Williamson Department of Computer Science University of Calgary
CPSC 641: Network Traffic Self-Similarity
Presentation transcript:

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 and Prediction of the Dynamic Behavior of Applications, Hosts, and Networks

3 Overview What is Self Similarity? Ethernet Traffic is Self-Similar Source of Self Similarity Implications of Self Similarity

4 Background Network traffic did not obey Poisson assumptions used in queuing analysis This paper, for the first time, provided an explanation and a systematic approach to modeling realistic data traffic patterns Sparked off research around the globe:  Results show self-similarity in ATM traffic, compressed digital video streams, and Web Traffic

5 Why is Self-Similarity Important? In this paper, Ethernet traffic has been identified as being self-similar. Models like Poisson are not able to capture the self-similarity property. This leads to inaccurate performance evaluation

Section 1: What is Self-Similarity ?

7 What is Self-Similarity? Self-similarity describes the phenomenon where a certain property of an object is preserved with respect to scaling in space and/or time. If an object is self-similar, its parts, when magnified, resemble the shape of the whole. In case of stochastic objects like time-series, self-similarity is used in the distributional sense

8 Intuition of Self-Similarity Something “feels the same” regardless of scale (also called fractals)

9

10

11

12 Self-Similarity in Traffic Measurement ( Ⅰ ) Traffic Measurement

13 Pictorial View of Self-Similarity

14 The Famous Data 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.

15

16 Why is Self-Similarity Important? 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 which, when applied to a huge network like the Internet, can lead to huge financial losses.

17 Problems with Current Models Current modeling shows that as the number of sources (Ethernet users) increases, the traffic becomes smoother and smoother Analysis shows that the traffic tends to become less smooth and more bursty as the number of active sources increases

18 Consequences of Self-Similarity Traffic has similar statistical properties at a range of timescales: ms, secs, mins, hrs, days Merging of traffic (as in a statistical multiplexer) does not result in smoothing of traffic Bursty Data Streams Aggregation Bursty Aggregate Streams

19 Problems with Current Models Cont.’d Were traffic to follow a Poisson or Markovian arrival process, it would have a characteristic burst length which would tend to be smoothed by averaging over a long enough time scale. Rather, measurements of real traffic indicate that significant traffic variance (burstiness) is present on a wide range of time scales

20 Pictorial View of Current Modeling

21 Side-by-side View

22 Definitions and Properties 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

23 Definitions and Properties Cont.’d low, medium, and high traffic hours as traffic increases, the Hurst parameter increases  i.e., traffic becomes more self-similar

24 Self-Similarity in Traffic Measurement ( Ⅱ ) Network Traffic

25 Properties of Self Similarity 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. E.g. X (1) = 4,12,34,2,-6,18,21,35 Then X (2) =8,18,6,28 X (4) =13,17 Mean , variance  2 Autocorrelation Function r(k) ~ k -b, where 0 < b < 1. Suppose that r(k)  k -β, 0<β<1

26 Auto-correlation Definition 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  

27 Traditional Models This is in contrast to traditional models Correlation structures of their aggregated processes degenerate as m   i.e. r (m) (k)  0 as m  for k = 1,2,3,... Example:  Poisson Distribution  Self-Similar Distribution

28

29 Long Range Dependence 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 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

30 Intuition Short-range processes:  Exponential Decay of autocorrelations, i.e.:  r(k) ~ p k, as k  , 0 < p < 1  Summation is finite Non-summability is an important property  Guarantees non-degenerate correlation structure of the aggregated processes X (m) as m  

31 The Measure of Self-Similarity Hurst Parameter H, 0.5 < H < 1 Three approaches to estimate H (Based on properties of self-similar processes)  Variance Analysis of aggregated processes  Analysis of Rescaled Range (R/S) statistic for different block sizes  A Whittle Estimator

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

33

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

35 Example 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 R/S = 1/S*[8-(-1)] = 9/S

36 The Hurst Effect 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 studies the 800-year record of flooding along the Nile river.  (yearly minimum water level)  Finds long-range dependence.

37

38 Whittle Estimator Provides a confidence interval Property: Any long range dependent process approaches FGN, when aggregated to a certain level Test the aggregated observations to ensure that it has converged to the normal distribution

39 Self Similarity X is exactly second-order self-similar with Hurst parameter H (= 1- β/2) if for all m,  Var(X (m) ) =  2 m -β, and  r (m) (k) = r(k), k  0 X is [asymptotically] second-order self-similar if the above holds when [ r (m) (k)  r(k), m  ∞ 

40 Modeling Self-Similarity Fractional Gaussian noise (FGN)  Gaussian process with mean , variance  2, and  Autocorrelation function r(k)=(|k+1| 2H -|k| 2H +|k-1| 2H ), k>0  Exactly second-order self-similar with 0.5<H<1 Fractional ARIMA(p,d,q)  Asymptotically second-order self-similar with H=d+0.5 where 0<d<0.5 Discrete time M/G/  input model  Service time X given by heavy tail distribution (i.e. E[x]<  2=  Example : Pareto distribution P(X>k) k -α, 1< α<2  N = {N t,t=1,2,…} is self-similar with H=(3- α)/2 where N t denotes # of members being serviced at time t

Section 2: Ethernet Traffic is Self-Similar

42 Plots Showing Self-Similarity ( Ⅰ ) H=0.5 H=1 Estimate H  0.8

43 Plots Showing Self-Similarity ( Ⅱ ) Higher Traffic, Higher H High Traffic Mid Traffic Low Traffic 1.3%-10.4% 3.4%-18.4% 5.0%-30.7%

44 Observation shows “contrary to Poisson”  Network UtilizationH  As we shall see shortly, H measures traffic burstiness As number of Ethernet users increases, the resulting aggregate traffic becomes burstier instead of smoother H : A Function of Network Utilization

45 Difference in low traffic H values 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

46 H : Measuring “Burstiness” Intuitive explanation using M/G/  Model  As α  1, service time is more variable, easier to generate burst  Increasing H ! Wrong way to measure “burstiness” of self- similar process  Peak-to-mean ratio  Coefficient of variation (for interarrival times)

47 Summary 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

48 Discussions How to explain self-similarity ?  Heavy tailed file sizes How this would impact existing performance?  Limited effectiveness of buffering  Effectiveness of FEC How to adapt to self-similarity?  Prediction  Adaptive FEC

49

50

Thanks !