NS2 HOME DESCRIPTION ANALYSIS CODE REPORT  
 
Network Simulation Project Demostration - Description

DESCRIPTION OF EXPERIMENTS

TCP may perform poorly over 802.11. We plan to ecalute TCP over 802.11. though multiple sets of experiments over the following topology, see Figure 1. We assume that the nodes are 150m apart and that they are equipped with IEEE 802.11b cards. To make things simple, we let the node do not move.

 



                              Figure 1. Wireless network node topology

 

we will investigate the impact of the flow control window with five typical different size. With each size, we will run some times and calculate the average throughput of specific link based on the 6-nodes chain.

Set of experiments A: Run each the following experiments ONCE for each of the following flow control windows: 1 packet (stop and wait), 2 packets, 4 packets, 8 packets, and 16 packets.

Experiment 1:  Establish a TCP connection between node #1 (sender) and node #2. Measure the throughput.

Experiment 2:  Establish a TCP connection between node #1 (sender) and node #3. Measure the throughput.

Experiment 3:  Establish a TCP connection between node #1 (sender) and node #4. Measure the throughput.

Experiment 4:  Establish a TCP connection between node #1 (sender) and node #5. Measure the throughput.

Experiment 5:  Establish a TCP connection between node #1 (sender) and node #6. Measure the throughput.

 

Set of experiments B: Start a TCP connection at a random time (between 0 and 0.5s) between node # 3 and node #4. Run each the five above experiments 30 times for the  five different values (1, 2, 4, 8, and 16) of the flow control window. With node #3 and node #4 connected all the time; we repeat the set of experiments A. See Figure 2 for the illustration of set of experiments B.

 

 

                             Figure 2. Network node topology of experiments B

 

Set of experiments C: Start one TCP connection at a random time (between 0 and 0.5s) between the following pairs: node #1 and node #6, node #2 and node#5a ,nd node # 3 and node #4. Run this experiment 30 times for the two topologies for the five different values (1, 2, 4, 8, and 16) of the flow control window.

 

                          Figure 3. Network node topology of experiments C

In the topology of experiments B and C, the above red line means that connections start with a random time at the same time. The bottom lines mean that in one subset of experiment, only one of these is connected with a start time of 0.

 

The duration of each run is 100 seconds. For each subset experiment, we run 30 times under the given flow control window, and calculate the average throughput.