| Author | Ganlaya Wuthisin |
| Call Number | AIT RSPR no. TC-01-06 |
| Subject(s) | TCP/IP (Computer network protocol)
|
| Note | A research study submitted in partial fulfillment of the requirements for the degree of
Master of Engineering, School of Advanced Technologies |
| Publisher | Asian Institute of Technology |
| Series Statement | Research studies project report ; no. TC-01-06 |
| Abstract | Recently, many studies have proposed that the Internet traffic dynamics resulting
from interaction among users, applications, and protocols in real networks is best
represented with the concept of Fractal traffic based on Self-Similar Processes. This
research studies the application-level fractal traffic, originating from a source that exhibits
fractal characteristics over a wide range of time and frequency scale without any
interaction with the network.
The application-level fractal traffic is generated by OPNET Modeler 8.0
simulator, which generates ON-OFF processes according to PowON-PowOFF model.
This research considers the aggregated traffic of I to 5 ON-OFF processes, Hurst
parameter from 0.65 to 0.95, Fractal Onset Timescale (FOTS) from 0.0001 to 0.01,
average ON processes rate 100-350 pkt/s, Mean of pareto and exponential packet size
distribution of ON processes from 4000 to 6000 bits and Standard deviation of pareto
distribution from 6x103 to 6x107
• All these parameters represent different characteristic of
aggregated traffic.
Simulation results show the effect of all parameters to loss probability, buffer
usage and delay when transmission capacity is constant at 2 Mbit/s with point-to-point
configuration. The FOTS is defined as the time scale from which fractal characteristics
begin to show up, which is used with the Hurst parameter. The high Hurst parameter and
low FOTS present the highly bursty aggregated traffic. The size of offered traffic is
determined by the source activity ratio and average arrival rate, which are calculated from
the number of ON-OFF processes and ON processes rate. The high source activity ratio
and average arrival rate present high offered traffic.
The results present that increase in Mean of packet size distribution increases
offered traffic then the loss probability, buffer usage and delay increase. The increase of
standard deviation with finite value has no effect on loss probability, buffer usage and
delay. |
| Year | 2001 |
| Corresponding Series Added Entry | Asian Institute of Technology. Research studies project report ; no. TC-01-06 |
| Type | Research Study Project Report (RSPR) |
| School | School of Advanced Technologies (SAT) |
| Department | Department of Information and Communications Technologies (DICT) |
| Academic Program/FoS | Telecommunications (TC) |
| Chairperson(s) | Erke, Tapio; |
| Examination Committee(s) | Ahmed, Kazi M;Kanchana Kanchanasut ; |
| Scholarship Donor(s) | - |
| Degree | Research Studies Project Report (M. Eng.) - Asian Institute of Technology, 2001 |