Lab3 Redes Opnet

Rick Wagoner
IET680 – Lab 03

This laboratory assignment is concerned with switched LANs. We are mainly concerned with the performance differences between a LAN using hubs and a LAN using switches.

We prepared two simulation scenarios. The first scenario, named OnlyHub, had 16 workstations connected through a single hub to create the LAN. The second scenario, named HubAndSwitch, had 1switch connected to two hubs that were each connected to 8 workstations. We can see from the following graph that the traffic sent from the workstations in both scenarios is practically identical.



Next we can look at the traffic that was received by the workstations in both scenarios. We can notice that the HubAndSwitch scenario showed a better throughput




Next we will look at thedelay in both scenarios. By reviewing this data we can see that we had practically no delay on the HubAndSwitch scenario but encountered a quick increase in the delay on the OnlyHub scenario before leveling off.



Finally we will look at the collision count under both scenarios. We can see the collision count for the OnlyHub scenario in the blue. This indicates a steep increase incollisions before leveling off around 2400 collisions per second. We can also see Hub1 in red and Hub2 in green in the SwitchAndHub scenario. Both hubs again show a steep increase but they then level off around 900 collisions per second.



We can see from the first three graphs that the throughput and delay are affected by the addition of a switch to the network. This is a prime example of howa switch and hub differ in operation. A hub receives a signal on an incoming line and then simply sends the same signal back out on all other lines. In our scenario HubOnly we received a signal in from Node 0. The hub simply sends this same signal back out on the remaining 15 lines. In this instance we increased our network traffic 15 times. This same occurrence is happening on every otherline as data is sent from the other nodes. This increase in network traffic increases congestion on the network which in turn reduces throughput as we tested and documented in our first two laboratory simulations. Scenario SwitchAndHub shows a better throughput because of the way the switch works. A switch only sends a received signal out on the line which connects to the destination of thetransmission. So in essence we are only doubling our traffic in this case. By reducing our traffic we can improve throughput.

Delay is affected even more by the different scenarios. Due to the increased traffic volume in the scenario HubOnly we encounter a larger delay. This again is a verification of the simulation we performed in Labs 1 and 2. In the SwitchAndHub scenario there is not as muchtotal traffic which lowers the total network delay.

Looking at our collisions we can also see a huge difference between the two scenarios. In the first scenario every node connected to the hub is part of the same collision domain. Since all the traffic in this scenario is in a single collision domain we encounter a much larger volume of total traffic and in turn a large number of collisions.The number of collisions appears to be directly proportional to the volume of the traffic on the network. As the traffic increases the number of collisions increases and once the network levels off at a peak traffic level the collisions also level off at a stable value. In the SwitchAndHub scenario we still have a large number of collisions but it is reduced because we now have two separatecollision domains (Hub1 and Hub2) and both are smaller in size generating less total traffic. In our SwitchAndHub scenario we are using the switch to reduce the number of nodes included in each collision domain. A switch in general makes every input line and every output line an individual collision domain thus collisions are greatly reduced. Hub1 and it’s nodes are one collision domain and...