You are on page 1of 4

Experiment 4

Objective:-

To generate non-Poisson distributed traffic, for the given value of mean and
peakedness factor.

Theory:-

A realistic traffic model should accommodate smooth and peaked traffic also in
addition to regular Poisson traffic. These can occur as the effect of internetworking
among different networks. The following model was used to generate non- Poisson
traffic.

Source 1

λ
1
Store the numbers Count the number
in buffer in each division.
λ
2 Random Non-Poisson
interrupt traffic

Source 2

Input parameters: -

Parameter Description
λ 1 Mean arrival rate of source 1.
λ 2 Mean arrival rate of source 2.
Random interrupt Value between 0 and 1 to switch between sources.
Number Number of values such as 10,000

Flowchart: -

A non–Poisson process similar to switched Poisson process (SPP) has to be generated.


Here two Poisson sources with mean arrival rates λ 1 and λ 2 respectively are taken. As
shown in the flowchart, generated values are mixed to get the non- Poisson traffic.
λ 1, λ 2, random interrupt will be chosen arbitrarily. For getting required mean
and peakedness factor, the input parameters λ 1, λ 2, random interrupt have to be varied
randomly and after large number of simulations, required λ 1, λ 2, random interrupt can
be obtained.

Start

Set Random interrupt


between 0 and 1.

Generate uniform Random


number between 0 and 1.

No
Yes Interrupt
<
Random number

Generate exponential
distributed value using Generate exponential
mean (1/λ 1) distributed value using
mean (1/λ 2)

Store the value in a


buffer and generate
10,000 such values.

Count the number in each division


and store them. These numbers
follow non- Poisson distribution.

Stop
Algorithm: -

Nonpoisson ( )
//Generates Non Poisson traffic.
//λ 1→ Mean arrival rate of source 1.
//λ 2→ Mean arrival rate of source 2.
//RI→ Random number between 0 and 1.
//RI→ Uniform distributed Random number between 0 and 1.
//RF→ Reference period.

START:
1.Select λ 1, λ 2, RI.
2.num←0;
3. while (num<10,000)
{
Generate a RN.
If (RN <RI) generate a exponential distributed value with mean (1/λ 1)
Else generate a exponential distributed value with mean (1/λ 2)
Num+=1;
}
4. RF←1.0;
5. Count the number of values in each RF and store that number.
6. Calculate mean and peakedness factor for these values.
7. If not matches with the desired values, select λ 1, λ 2, RI randomly again and
Repeat the procedure {step 2 to step 6}.
END.

Output:-
Traffic following non-Poisson distribution.
Results:-
Tutorial problem: -
1. Generate non- Poisson distributed values for a mean arrival rate of
0.067 calls/sec and for a peakedness factor of 0.9,1.1,1.4.

You might also like