광역 네트워크 트래픽의 장거리 상관관계와 노이즈 = Long-Range Dependence and Noise in a Wide Area Network Traffic

다국어 초록 (Multilingual Abstract)

In this paper, we examine a long-range dependence in an active measurement of a network traffic which has been a well known characteristic from analyses of a passive network traffic measurement. To this end, we utilize RTT(Round Trip Time), which is a typical active measurement measured by PingER project, and perform a relevant analysis to a time series of both RTT and its volatilities. The RTT time series exhibits a long-range dependence or a noise. The volatilities, defined as a higher-order variation, follow a log-normal distribution. Furthermore, volatilities show a long-range dependence in relatively short time intervals, and a long-range dependence and/or noise in long time intervals. From this study, we find that the long-range dependence is a characteristic of not only a passive traffic measurement but also an active measurement of network traffic such as RTT. From these findings, we can infer that the long-range dependence is a characteristic of network traffic independent of a type of measurements. In particular, an active measurement exhibits a noise which cannot be usually found in a passive measurement.

국문 초록 (Abstract)

본 논문에서는 네트워크 트래픽의 수동적 측정치 분석을 통해 잘 알려진 장거리 상관관계가 광역 네트워크의 능동적 측정치에도 존재하는지 여부를 관련 분석법을 통하여 검정하고자 한다. 이를 위하여 PingER 프로젝트를 통하여 측정된 광역 네트워크 트래픽의 대표적인 능동적 측정치인 RTT(Round Trip Time)와 RTT의 변동성 시계열 데이터에 대하여 분석을 수행하였다. RTT 시계열 데이터는 장거리 상관관계 혹은 노이즈의 특성을 보였으며, RTT의 고차원 변화량으로 정의된 변동성은 로그정규분포를 따르며 변동성에 대한 장거리 상관관계는 고려하는 시간 간격이 짧은 경우 장거리 상관관계를 보이고, 시간 간격이 긴 경우에는 장거리 상관관계 혹은 노이즈를 따름을 밝혔다. 본 연구를 통해 볼 때 장거리 상관관계는 비단 패킷 도착의 시간 간격 등과 같은 수동적 측정뿐만 아니라 RTT와 같은 능동적 측정에서도 나타나는 특징이며, 특히 능동적 측정에는 수동적 측정에는 잘 나타나지 않는 노이즈 특성이 존재함을 밝혔다.

참고문헌 (Reference)

1 "http://www-iepm.slac.stanford.edu/pinger/"

2 Paxson, V, "Wide-Area traffic: The failure of Poisson modeling" 3 (3): 226-244, 1995

3 Müller, U, "Volatilities of different time resolutions - Analyzing the dynamics of market components" 4 : 213-239, 1997

4 Leland, W, "The Self-similar nature of Ethernet traffic" 2 (2): 1-15, 1994

5 Beran, J, "Statistics for Long-memory Processes" Chapman & Hall/CRC 1994

6 Liu, Y, "Statistical properties of the volatility of price fluctuations" 62 (62): 1390-1400, 1999

7 Ribeiro, V, "Small- time scaling behavior of Internet backbone traffic" 48 : 315-334, 2005

8 Willinger, W, "Self-similarity through high-variability: statistical analysis of Ethernet LAN traffic at the source level" 5 (5): 71-86, 1997

9 Bak, P, "Self- organized criticality: An explanation of the 1/f noise" 59 (59): 381-384, 1987

10 Peng, C, "Quantification of scaling exponents and crossover phenomena in nonstationary heartbeat time series" 5 (5): 82-87, 1995

11 Scherrer, A, "Non-Gaussian and Long Memory Statistical Characterizations for Internet Traffic with Anomalies" 4 : 56-70, 2007

12 Veitch, D, "Multifractality in TCP/IP traffic: the case against" 48 : 293-313, 2004

13 Peng, C, "Mosaic organization of DNA nucleotides" 40 (40): 1685-1689, 1994

14 Karagiannis, T, "Long-Range Dependence: Ten Years of Internet Traffic Modeling" 8 (8): 57-64, 2004

15 Hogg, R, "Introduction to mathematical statistics" Macmillan Publishing Co 1978

16 Claffy, K, "Internet measurement and data analysis: topology, workload, performance, and routing statistics" 1999

17 Chang-Yong Lee, "Higher-order Correlations in Data Network Traffic" 한국물리학회 45 (45): 1664-1670, 2004

18 Paxson,V, "Growth trends in wide-area TCP connections" 8 (8): 8-17, 1994

19 Bollerslev,T, "Generalized autoregressive conditional heteroskedasticity" 31 (31): 307-327, 1986

20 Takayasu, M, "Dynamic phase transition observed in the Internet traffic flow" 277 : 248-255, 2000

21 Xiao-Yan, Z, "Detrended Fluctuation Analysis of Traffic Data" 24 : 2142-2145, 2007

22 Nakayama,A, "Critical behaviors and self-similarity in models of computer network traffic" 293 : 285-296, 2001

23 Lee.C, "Characteristics of the volatility in the Korea composite stock price index" 388 : 3837-3850, 2009

24 Engle,R, "Autoregressive Conditional Heteroscedasticity with Estimates of the Variance of United Kingdom Inflation" 50 (50): 987-1007, 1982