Opnet Simulation ppt

1
Simulation with OPNET for Verification
2
System Architecture

• Campus model wireless Lan
• Transmission medium is air.
• 1 server, 11 Nodes Star Topology
• Each node produces Email, fttp
• http and voice traffic according
• to the specified distribution.
• Center node relays all traffic.
• Application configuration
• which apps are going to be
• simulated.
• Profile configuration
• creates a profile that uses these
• apps configured in application
• configuration.

3
Server Settings

• Mostly default values
• Created profile is used for
• Application supported services
• Name and model are defined
• Detailed configuration are made
• in Lan Parameters menu

4
Lan Parameters

• Data Rate is changed to 2 Mbps.
• Access Point Functionality is enabled for server

5
Workstation Settings

• Mostly default values
• Created profile is used for
• Application supported profiles
• Name and model are defined
• Detailed configuration are made
• in Lan Parameters menu

6
Simulation Configuration

• Simulation duration is 1 hr.
• Wireless Lan range is 300 mt.
• Also the required reports are
• chosen.
• Throughput
• Delay
• Lan Load

7
Running Simulation

• It took 23 seconds to run the
• simulation.
• The simulation was run twice
• first with Uniform Distribution for
• message arrivals
• second with Poisson Distribution

8
Simulation Reports for Server (Uniform Dist.)
Uniform Distribution min Outcome 0 max
Outcome 1000
Load (bits/sec) Min 0 Max 6000 Mean
4200
Delay (sec) Min 0.0004 Max 0.0024
Mean 0.0014
Throughput (bits/sec) Min 0 Max 5000
Mean 3900
9
Simulation Reports for Workstation 1 (Uniform
Dist.)
Load (bits/sec) Min 0 Max 400 Mean
330
Delay (sec) Min 0.0004 Max 0.0007
Mean 0.0006
Throughput (bits/sec) Min 0 Max 1200
Mean 990
10
Simulation Reports for Server (Poisson Dist.)
Poisson Distribution mean Outcome 500
Load (bits/sec) Min 0 Max 3800 Mean
3300
Delay (sec) Min 0 Max 0.0022 Mean
0.0015
Throughput (bits/sec) Min 0 Max 2100
Mean 1400
11
Simulation Reports for Workstation 1 (Poisson
Dist.)
Load (bits/sec) Min 0 Max 290 Mean
220
Delay (sec) Min 0 Max 0.0007 Mean
0.0006
Throughput (bits/sec) Min 0 Max 800
Mean 690
12
Comparison Analytical versus Simulation
l 300 mt. Lm 8000 bits ? 40 bit/sec
13
Two teams structure with command center

• Two teams and command center
• system.
• Both subnets are connected to
• the command center (router).
• Center router acts as a relay
• station between the subnets.
• Same distributions are used to
• simulate the traffic.
• All system settings are same with
• the previous model.

14
Inside the Subnet

• Subnets are identical.
• 1 server (access point), 11 nodes
• Center node relays all traffic
• between the nodes and acts as
• a gateway for the subnet.
• Each node produces Email, fttp
• http and voice traffic according
• to the specified distribution.
• Distances between nodes and the
• center is 300 m., data rate 2Mbps.

15
Access Point 1
Load (bits/sec) Min 0 Max 350 Mean
240
Delay (sec) Min 0 Max 0.0008 Mean
0.0005
Throughput (bits/sec) Min 0 Max 120
Mean 90
16
Access Point 2
Load (bits/sec) Min 0 Max 120 Mean
90
Delay (sec) Min 0 Max 0.00025 Mean
0.0002
Throughput (bits/sec) Min 0 Max 100
Mean 80
17
Server 1
Load (bits/sec) Min 0 Max 120 Mean
90
Delay (sec) Min 0.00008 Max 0.00022
Mean 0.00018
8
Throughput (bits/sec) Min 0 Max 170
Mean 110
18
Server 2
Load (bits/sec) Min 0 Max 70 Mean
50
Delay (sec) Min 0 Max 0.00022 Mean
0.00018
Throughput (bits/sec) Min 0 Max 100
Mean 60
19
Conclusions

• According to the analytical method, the mean
  total delay of the system is 8 ms.
• This value is verified as 8 ms, with the
  simulation model.
• The mean total delay is mostly affected by the
  message size and the arrival rate
• The other variable is the distances between the
  nodes and the central node.
• We have assumed that the distances are
  constant. To get more realistic results
• the distances can be represented with a
  distribution.

20
Conclusions

• This project was a good hands-on experience to
  understand the concepts
• we covered in the course
• The simulation tool that we have used, is very
  powerful, however we have
• spent a long time to make it work.
• Opnet Modeler is not user-friendly. It requires
  a special training.