Self-similar traffic generation
View Patent ↗There is disclosed apparatus and methods for mimicking self-similar traffic in a telecommunications network. One ore more self-similar streams of data units may be generated according to values of selected parameters for the streams, such as the Hurst parameter. Values may be pre-calculated and stored in a table in a memory. To generate the stream, a traffic distribution is generated based on the values of the parameters, and through retrieval of the values from the table.
1. A method for mimicking self-similar traffic in a telecommunications network, the method comprising
providing a table of distributed sequences for self-similar traffic
providing values for at least one self-similarity parameter for plural streams, the self-similarity parameters including a Hurst parameter
selecting at least one of the streams
generating a self-similar traffic distribution for the selected streams based on the values of the self similarity parameters for the selected streams
wherein the self-similar traffic distribution is further based upon distributed sequences retrieved from the table based upon the value of the Hurst parameter for the selected streams
wherein the distributed sequences are described by a formula having a shape parameter and a minimum value for the respective distributed sequence and a uniformly distributed random variable
wherein the value for the shape parameter is fixed and the minimum value for the distributed sequences is fixed
wherein the table comprises logarithms of values of the uniformly distributed random variable.
2. The method for mimicking self-similar traffic in a telecommunications network of claim 1 , wherein the distributed sequences are Pareto-distributed sequences.
3. The method for mimicking self-similar traffic in a telecommunications network of claim 1 , wherein the distributed sequences are truncated-value Pareto distributions.
4. The method for mimicking self-similar traffic in a telecommunications network of claim 1 , wherein the distributed sequences comprise competing flows.
5. The method for mimicking self-similar traffic in a telecommunications network of claim 4 wherein the distributed sequences each comprise a single ON/OFF process.
6. The method for mimicking self-similar traffic in a telecommunications network of claim 1 , wherein the traffic distribution comprises a series of ON and OFF periods.
7. The method for mimicking self-similar traffic in a telecommunications network of claim 6 , wherein only one of the ON and OFF periods has a heavy-tailed distribution.
8. The method for mimicking self-similar traffic in a telecommunications network of claim 7 , wherein the heavy-tailed distribution is a Pareto distribution.
9. The method for mimicking self-similar traffic in a telecommunications network of claim 1 wherein the distributed sequences comprise superposition of plural independent and probabilistically identical fractal renewal processes.
10. The method for mimicking self-similar traffic in a telecommunications network of claim 1 , wherein the self-similarity parameters include an average arrival time and a fractal onset time.
11. The method for mimicking self-similar traffic in a telecommunications network of claim 10 wherein the values of the average arrival time and the fractal onset time are used for retrieving the distributed sequences from the table in conjunction with the Hurst parameter.
12. The method for mimicking self-similar traffic in a telecommunications network of claim 1 further comprising generating the selected streams from the self-similar traffic distribution.
13. The method for mimicking self-similar traffic in a telecommunications network of claim 12 further comprising applying the generated streams of self-similar traffic to a portion of a telecommunications network.
14. The method for mimicking self-similar traffic in a telecommunications network of claim 13 further comprising simulating a behavior of the portion of the telecommunications network based on the applied streams of self-similar traffic.
15. A method for mimicking self-similar traffic in a telecommunications network, the method comprising
providing values for at least one self-similarity parameter for plural streams, the self-similarity parameters including a Hurst parameter
selecting at least one of the streams
generating a self-similar traffic distribution for the selected streams based on the values of the self similarity parameters for the selected streams, wherein the self-similar traffic distribution is further based upon distributed sequences retrieved from a memory based upon the value of the Hurst parameter for the selected streams
wherein the memory stores distributed sequences for self-similar traffic correlated to a range of values for the Hurst parameter
wherein the distributed sequences are described by a formula having a shape parameter and a minimum value for the respective distributed sequence and a uniformly distributed random variable
wherein the value for the shape parameter is fixed and the minimum value for the distributed sequences is fixed
wherein the memory stores logarithms of values of the uniformly distributed random variable.
16. The method for mimicking self-similar traffic in a telecommunications network of claim 15 , wherein the distributed sequences are Pareto-distributed sequences.
17. The method for mimicking self-similar traffic in a telecommunications network of claim 15 , wherein the distributed sequences are truncated-value Pareto distributions.
18. The method for mimicking self-similar traffic in a telecommunications network of claim 15 , wherein the distributed sequences comprise competing flows.
19. The method for mimicking self-similar traffic in a telecommunications network of claim 18 wherein the distributed sequences each comprise a single ON/OFF process.
20. The method for mimicking self-similar traffic in a telecommunications network of claim 15 , wherein the traffic distribution comprises a series of ON and OFF periods.
21. The method for mimicking self-similar traffic in a telecommunications network of claim 20 , wherein only one of the ON and OFF periods has a heavy-tailed distribution.
22. The method for mimicking self-similar traffic in a telecommunications network of claim 21 , wherein the heavy-tailed distribution is a Pareto distribution.
23. The method for mimicking self-similar traffic in a telecommunications network of claim 15 wherein the distributed sequences comprise superposition of plural independent and probabilistically identical fractal renewal processes.
24. The method for mimicking self-similar traffic in a telecommunications network of claim 15 , wherein the self-similarity parameters include an average arrival time and a fractal onset time.
25. The method for mimicking self-similar traffic in a telecommunications network of claim 24 wherein the values of the average arrival time and the fractal onset time are used for retrieving the distributed sequences from the memory in conjunction with the Hurst parameter.
26. The method for mimicking self-similar traffic in a telecommunications network of claim 15 further comprising generating the selected streams from the self-similar traffic distribution.
27. The method for mimicking self-similar traffic in a telecommunications network of claim 26 further comprising applying the generated streams of self-similar traffic to a portion of a telecommunications network.
28. The method for mimicking self-similar traffic in a telecommunications network of claim 27 further comprising simulating a behavior of the portion of the telecommunications network based on the applied streams of self-similar traffic.
29. A method for mimicking self-similar traffic in a telecommunications network, the method comprising
providing values for at least one self-similarity parameter for plural streams, the self-similarity parameters including a Hurst parameter
selecting at least one of the streams
generating self-similar traffic distributions for the selected streams based the values of on the self similarity parameters for the selected streams, wherein
the self-similar traffic distribution is based upon distributed sequences described by a formula having a shape parameter, a minimum value for the respective distributed sequences and a uniformly distributed random variable
the shape parameter has a fixed value
the minimum value for the distributed sequences is fixed
the distributed sequences are obtained from logarithms of values of the uniformly distributed random variable.
30. The method for mimicking self-similar traffic in a telecommunications network of claim 29 further comprising
pre-storing the logarithms of values of the uniformly distributed random variable in a memory
in the generating step, obtaining the logarithms from the memory.
31. The method for mimicking self-similar traffic in a telecommunications network of claim 29 further comprising
pre-storing the distributed sequences in a memory
in the generating step, obtaining the distributed sequences from the memory.
32. The method for mimicking self-similar traffic in a telecommunications network of claim 29 , wherein the distributed sequences are Pareto-distributed sequences.
33. The method for mimicking self-similar traffic in a telecommunications network of claim 29 , wherein the distributed sequences are truncated-value Pareto distributions.
34. The method for mimicking self-similar traffic in a telecommunications network of claim 29 , wherein the distributed sequences comprise competing flows.
35. The method for mimicking self-similar traffic in a telecommunications network of claim 34 wherein the distributed sequences each comprise a single ON/OFF process.
36. The method for mimicking self-similar traffic in a telecommunications network of claim 29 , wherein the traffic distribution comprises a series of ON and OFF periods.
37. The method for mimicking self-similar traffic in a telecommunications network of claim 36 , wherein only one of the ON and OFF periods has a heavy-tailed distribution.
38. The method for mimicking self-similar traffic in a telecommunications network of claim 37 , wherein the heavy-tailed distribution is a Pareto distribution.
39. The method for mimicking self-similar traffic in a telecommunications network of claim 29 wherein the distributed sequences comprise superposition of plural independent and probabilistically identical fractal renewal processes.
40. The method for mimicking self-similar traffic in a telecommunications network of claim 29 , wherein the self-similarity parameters include an average arrival time and a fractal onset time.
41. The method for mimicking self-similar traffic in a telecommunications network of claim 40 wherein the values of the average arrival time and the fractal onset time are used for retrieving the distributed sequences from the memory in conjunction with the Hurst parameter.
42. The method for mimicking self-similar traffic in a telecommunications network of claim 29 further comprising generating the selected streams from the self-similar traffic distribution.
43. The method for mimicking self-similar traffic in a telecommunications network of claim 42 further comprising applying the generated streams of self-similar traffic to a portion of a telecommunications network.
44. The method for mimicking self-similar traffic in a telecommunications network of claim 43 further comprising simulating a behavior of the portion of the telecommunications network based on the applied streams of self-similar traffic.
45. An apparatus for mimicking self-similar traffic in a telecommunications network, the apparatus comprising
a buffer to receive values for at least one self-similarity parameter for plural streams, the self-similarity parameters including a Hurst parameter
an arbitrator coupled to the buffer to select at least one of the streams
a memory storing a table of distributed sequences for self-similar traffic
a distribution calculator coupled to the memory, the distribution calculator to generate a self-similar traffic distribution for the selected streams based on the values of the self similarity parameters for the selected streams
wherein the self-similar traffic distribution is further based upon distributed sequences retrieved from the table based upon the value of the Hurst parameter for the selected streams
wherein the distributed sequences are described by a formula having a shape parameter and a minimum value for the respective distributed sequence and a uniformly distributed random variable
wherein the value for the shape parameter is fixed and the minimum value for the distributed sequences is fixed
wherein the table comprises logarithms of values of the uniformly distributed random variable
a scheduler coupled to the distribution calculator to schedule transmission of data units for the selected streams.
46. The apparatus for mimicking self-similar traffic in a telecommunications network of claim 45 , wherein the distributed sequences are Pareto-distributed sequences.
47. The apparatus for mimicking self-similar traffic in a telecommunications network of claim 45 , wherein the distributed sequences are truncated-value Pareto distributions.
48. The apparatus for mimicking self-similar traffic in a telecommunications network of claim 45 , wherein the distributed sequences comprise competing flows.
49. The apparatus for mimicking self-similar traffic in a telecommunications network of claim 48 wherein the distributed sequences each comprise a single ON/OFF process.
50. The apparatus for mimicking self-similar traffic in a telecommunications network of claim 45 , wherein the traffic distribution comprises a series of ON and OFF periods.
51. The apparatus for mimicking self-similar traffic in a telecommunications network of claim 50 , wherein only one of the ON and OFF periods has a heavy-tailed distribution.
52. The apparatus for mimicking self-similar traffic in a telecommunications network of claim 51 , wherein the heavy-tailed distribution is a Pareto distribution.
53. The apparatus for mimicking self-similar traffic in a telecommunications network of claim 45 wherein the distributed sequences comprise superposition of plural independent and probabilistically identical fractal renewal processes.
54. The apparatus for mimicking self-similar traffic in a telecommunications network of claim 45 , wherein the self-similarity parameters include an average arrival time and a fractal onset time.
55. The apparatus for mimicking self-similar traffic in a telecommunications network of claim 54 wherein the values of the average arrival time and the fractal onset time are used for retrieving the distributed sequences from the table in conjunction with the Hurst parameter.
56. The apparatus for mimicking self-similar traffic in a telecommunications network of claim 45 further comprising a blaster unit to generate the selected streams from the self-similar traffic distribution.
57. The apparatus for mimicking self-similar traffic in a telecommunications network of claim 56 wherein the blaster unit includes the scheduler.
58. The apparatus for mimicking self-similar traffic in a telecommunications network of claim 56 further comprising a transmit engine to apply the generated streams of self-similar traffic to a portion of a telecommunications network.
59. A network card comprising the apparatus of claim 45 .
60. A traffic generator comprising the apparatus of claim 45 .
61. An apparatus for mimicking self-similar traffic in a telecommunications network, the apparatus comprising
a buffer to receive values for at least one self-similarity parameter for plural streams, the self-similarity parameters including a Hurst parameter
an arbitrator coupled to the buffer to select at least one of the streams
a memory storing distributed sequences for self-similar traffic correlated to a range of values for the Hurst parameter
a distribution calculator coupled to the memory, the distribution calculator to generate a self-similar traffic distribution for the selected streams based on the values of the self similarity parameters for the selected streams
wherein the self-similar traffic distribution is further based upon distributed sequences retrieved from the memory based upon the value of the Hurst parameter for the selected streams
wherein the distributed sequences are described by a formula having a shape parameter and a minimum value for the respective distributed sequence and a uniformly distributed random variable
wherein the value for the shape parameter is fixed and the minimum value for the distributed sequences is fixed
wherein the memory stores logarithms of values of the uniformly distributed random variable
a scheduler coupled to the distribution calculator to schedule transmission of data units for the selected streams.
62. The apparatus for mimicking self-similar traffic in a telecommunications network of claim 61 , wherein the distributed sequences are Pareto-distributed sequences.
63. The apparatus for mimicking self-similar traffic in a telecommunications network of claim 61 , wherein the distributed sequences are truncated-value Pareto distributions.
64. The apparatus for mimicking self-similar traffic in a telecommunications network of claim 61 , wherein the distributed sequences comprise competing flows.
65. The apparatus for mimicking self-similar traffic in a telecommunications network of claim 61 wherein the distributed sequences each comprise a single ON/OFF process.
66. The apparatus for mimicking self-similar traffic in a telecommunications network of claim 61 , wherein the traffic distribution comprises a series of ON and OFF periods.
67. The apparatus for mimicking self-similar traffic in a telecommunications network of claim 66 , wherein only one of the ON and OFF periods has a heavy-tailed distribution.
68. The apparatus for mimicking self-similar traffic in a telecommunications network of claim 67 , wherein the heavy-tailed distribution is a Pareto distribution.
69. The apparatus for mimicking self-similar traffic in a telecommunications network of claim 61 wherein the distributed sequences comprise superposition of plural independent and probabilistically identical fractal renewal processes.
70. The apparatus for mimicking self-similar traffic in a telecommunications network of claim 61 , wherein the self-similarity parameters include an average arrival time and a fractal onset time.
71. The apparatus for mimicking self-similar traffic in a telecommunications network of claim 70 wherein the values of the average arrival time and the fractal onset time are used for retrieving the distributed sequences from the memory in conjunction with the Hurst parameter.
72. The apparatus for mimicking self-similar traffic in a telecommunications network of claim 61 further comprising a blaster unit to generate the selected streams from the self-similar traffic distribution.
73. The apparatus for mimicking self-similar traffic in a telecommunications network of claim 72 wherein the blaster unit includes the scheduler.
74. The apparatus for mimicking self-similar traffic in a telecommunications network of claim 72 further comprising a transmit engine to apply the generated streams of self-similar traffic to a portion of a telecommunications network.
75. A network card comprising the apparatus of claim 61 .
76. A traffic generator comprising the apparatus of claim 61 .
77. An apparatus for mimicking self-similar traffic in a telecommunications network, the apparatus comprising
a buffer to receive values for at least one self-similarity parameter for plural streams, the self-similarity parameters including a Hurst parameter
an arbitrator coupled to the buffer to select at least one of the streams
a memory storing distributed sequences for self-similar traffic correlated to a range of values for the Hurst parameter
a distribution calculator coupled to the memory, the distribution calculator to generate self-similar traffic distributions for the selected streams based on the values of the self similarity parameters for the selected streams, wherein
the self-similar traffic distribution is based upon distributed sequences described by a formula having a shape parameter, a minimum value for the respective distributed sequences and a uniformly distributed random variable
the shape parameter has a fixed value
the minimum value for the distributed sequences is fixed
the distributed sequences are obtained from logarithms of values of the uniformly distributed random variable
a scheduler coupled to the distribution calculator to schedule transmission of data units for the selected streams.
78. The apparatus of claim 77 wherein the memory stores the logarithms of values of the uniformly distributed random variable, and the distribution calculator obtains the logarithms from the memory.
79. The apparatus of claim 77 wherein the memory stores the distributed sequences, and the distribution calculator obtains the distributed sequences from the memory.
80. The apparatus for mimicking self-similar traffic in a telecommunications network of claim 77 , wherein the distributed sequences are Pareto-distributed sequences.
81. The apparatus for mimicking self-similar traffic in a telecommunications network of claim 77 , wherein the distributed sequences are truncated-value Pareto distributions.
82. The apparatus for mimicking self-similar traffic in a telecommunications network of claim 77 , wherein the distributed sequences comprise competing flows.
83. The apparatus for mimicking self-similar traffic in a telecommunications network of claim 82 wherein the distributed sequences each comprise a single ON/OFF process.
84. The apparatus for mimicking self-similar traffic in a telecommunications network of claim 77 , wherein the traffic distribution comprises a series of ON and OFF periods.
85. The apparatus for mimicking self-similar traffic in a telecommunications network of claim 84 , wherein only one of the ON and OFF periods has a heavy-tailed distribution.
86. The apparatus for mimicking self-similar traffic in a telecommunications network of claim 85 , wherein the heavy-tailed distribution is a Pareto distribution.
87. The apparatus for mimicking self-similar traffic in a telecommunications network of claim 77 wherein the distributed sequences comprise superposition of plural independent and probabilistically identical fractal renewal processes.
88. The apparatus for mimicking self-similar traffic in a telecommunications network of claim 77 , wherein the self-similarity parameters include an average arrival time and a fractal onset time.
89. The apparatus for mimicking self-similar traffic in a telecommunications network of claim 88 wherein the values of the average arrival time and the fractal onset time are used for retrieving the distributed sequences from the memory in conjunction with the Hurst parameter.
90. The apparatus for mimicking self-similar traffic in a telecommunications network of claim 77 further comprising a blaster unit to generate the selected streams from the self-similar traffic distribution.
91. The apparatus for mimicking self-similar traffic in a telecommunications network of claim 90 wherein the blaster unit includes the scheduler.
92. The apparatus for mimicking self-similar traffic in a telecommunications network of claim 90 further comprising a transmit engine to apply the generated streams of self-similar traffic to a portion of a telecommunications network.
93. A network card comprising the apparatus of claim 77 .
94. A traffic generator comprising the apparatus of claim 77 .
95. A computer-readable medium having computer readable program code embodied therein for mimicking self-similar traffic in a telecommunications network, the computer readable code which when executed by a computer causing the computer to perform:
receive values for at least one self-similarity parameter for plural streams, the self-similarity parameters including a Hurst parameter
select at least one of the streams
generate a self-similar traffic distribution for the selected streams based on the values of the self similarity parameters for the selected streams
wherein the self-similar traffic distribution is further based upon distributed sequences retrieved from a table based upon the value of the Hurst parameter for the selected streams
wherein the table stores distributed sequences for self-similar traffic
wherein the distributed sequences are described by a formula having a shape parameter and a minimum value for the respective distributed sequence and a uniformly distributed random variable
wherein the value for the shape parameter is fixed and the minimum value for the distributed sequences is fixed
wherein the table comprises logarithms of values of the uniformly distributed random variable.
96. The computer-readable medium of claim 95 , wherein the distributed sequences are Pareto-distributed sequences.
97. The computer-readable medium of claim 95 , wherein the distributed sequences are truncated-value Pareto distributions.
98. The computer-readable medium of claim 95 , wherein the distributed sequences comprise competing flows.
99. The computer-readable medium of claim 98 wherein the distributed sequences each comprise a single ON/OFF process.
100. The computer-readable medium of claim 95 , wherein the traffic distribution comprises a series of ON and OFF periods.
101. The computer-readable medium of claim 100 , wherein only one of the ON and OFF periods has a heavy-tailed distribution.
102. The computer-readable medium of claim 101 , wherein the heavy-tailed distribution is a Pareto distribution.
103. The computer-readable medium of claim 95 wherein the distributed sequences comprise superposition of plural independent and probabilistically identical fractal renewal processes.
104. The computer-readable medium of claim 95 , wherein the self-similarity parameters include an average arrival time and a fractal onset time.
105. The computer-readable medium of claim 104 , the computer readable code for causing a processor to use the values of the average arrival time and the fractal onset time to retrieve the distributed sequences from the table in conjunction with the Hurst parameter.
106. A computer-readable medium having computer readable program code embodied therein for mimicking self-similar traffic in a telecommunications network, the computer readable code which when executed by a computer causing the computer to perform:
receive values for at least one self-similarity parameter for plural streams, the self-similarity parameters including a Hurst parameter
select at least one of the streams
generate a self-similar traffic distribution for the selected streams based on the values of the self similarity parameters for the selected streams
wherein the self-similar traffic distribution is further based upon distributed sequences retrieved from a memory based upon the value of the Hurst parameter for the selected streams
wherein the memory stores distributed sequences for self-similar traffic correlated to a range of values for the Hurst parameter
wherein the distributed sequences are described by a formula having a shape parameter and a minimum value for the respective distributed sequence and a uniformly distributed random variable
wherein the value for the shape parameter is fixed and the minimum value for the distributed sequences is fixed
wherein the memory stores logarithms of values of the uniformly distributed random variable.
107. The computer-readable medium of claim 106 , wherein the distributed sequences are Pareto-distributed sequences.
108. The computer-readable medium of claim 106 , wherein the distributed sequences are truncated-value Pareto distributions.
109. The computer-readable medium of claim 106 , wherein the distributed sequences comprise competing flows.
110. The computer-readable medium of claim 106 wherein the distributed sequences each comprise a single ON/OFF process.
111. The computer-readable medium of claim 106 , wherein the traffic distribution comprises a series of ON and OFF periods.
112. The computer-readable medium of claim 111 , wherein only one of the ON and OFF periods has a heavy-tailed distribution.
113. The computer-readable medium of claim 112 , wherein the heavy-tailed distribution is a Pareto distribution.
114. The computer-readable medium of claim 106 wherein the distributed sequences comprise superposition of plural independent and probabilistically identical fractal renewal processes.
115. The computer-readable medium of claim 106 , wherein the self-similarity parameters include an average arrival time and a fractal onset time.
116. The computer-readable medium of claim 115 wherein the values of the average arrival time and the fractal onset time are used for retrieving the distributed sequences from the memory in conjunction with the Hurst parameter.
117. A computer-readable medium having computer readable program code embodied therein for mimicking self-similar traffic in a telecommunications network, the computer readable code which when executed by a computer causing the computer to perform:
receive values for at least one self-similarity parameter for plural streams, the self-similarity parameters including a Hurst parameter
select at least one of the streams
generate self-similar traffic distributions for the selected streams based on the values of the self similarity parameters for the selected streams, wherein
the self-similar traffic distribution is based upon distributed sequences described by a formula having a shape parameter, a minimum value for the respective distributed sequences and a uniformly distributed random variable
the shape parameter has a fixed value
the minimum value for the distributed sequences is fixed
the distributed sequences are obtained from logarithms of values of the uniformly distributed random variable.
118. The computer-readable medium of claim 117 further comprising
pre-storing the logarithms of values of the uniformly distributed random variable in a memory
in the generating step, obtaining the logarithms from the memory.
119. The computer-readable medium of claim 117 further comprising
pre-storing the distributed sequences in a memory
in the generating step, obtaining the distributed sequences from the memory.
120. The computer-readable medium of claim 117 , wherein the distributed sequences are Pareto-distributed sequences.
121. The computer-readable medium of claim 117 , wherein the distributed sequences are truncated-value Pareto distributions.
122. The computer-readable medium of claim 117 , wherein the distributed sequences comprise competing flows.
123. The computer-readable medium of claim 122 wherein the distributed sequences each comprise a single ON/OFF process.
124. The computer-readable medium of claim 117 , wherein the traffic distribution comprises a series of ON and OFF periods.
125. The computer-readable medium of claim 124 , wherein only one of the ON and OFF periods has a heavy-tailed distribution.
126. The computer-readable medium of claim 125 , wherein the heavy-tailed distribution is a Pareto distribution.
127. The computer-readable medium of claim 117 wherein the distributed sequences comprise superposition of plural independent and probabilistically identical fractal renewal processes.
128. The computer-readable medium of claim 117 , wherein the self-similarity parameters include an average arrival time and a fractal onset time.
129. The computer-readable medium of claim 128 , the computer readable code for causing a processor to use the values of the average arrival time and the fractal onset time to retrieve the distributed sequences from the memory in conjunction with the Hurst parameter.