Local Area Networks Interconnection

This lab is designed to demonstrate the implementation of switched local area networks. The simulation in this lab will help you examine the performance of different implementations of local area networks connected by switches and hubs.

A hub forwards the packet that arrives on any of its inputs on all the outputs regardless of the destination of the packet. Hub has only one collision domain.

On the other hand, a switch forwards incoming packets to one or more outputs depending on the destination(s) of the packets. Switches has separate collision domains for each connection.

Here we will study how the throughput and collision of packets in a switched network are affected by the configuration of the network and the types of switching devices that are used.

Implementation of Network

The implementation involves series of steps as follows:

We select the network topology as star and edit all the attributes as per the lab manual.

The next step is to configure the nodes of the network for setting the traffic generated by each station. The network consists of 16 nodes each of which is connected to a hub by a 10 Base-T Ethernet.

The statistics are chosen for the simulation. We select the delay(sec), traffic sent ( packets/sec), traffic received ( packets/sec) and the collision count.

The simulation is configured for the time duration of 2 minutes.

The next step involves duplicating the scenario in order to have two hubs connected by one switch.

Finally, the simulation is run and the results are compared for their performance based on the statistics chosen for simulation.

Lab Results:

We have experimented with two scenarios i.e.

Scenario1: Only hub

Scenario2: Two hubs and a switch.

The statistics chosen for comparing the above mentioned scenarios are

Ethernet Delay (sec)

Traffic sent (packets/sec)

Traffic received (packets/sec)

Collision count.

As per the implementation steps followed, we received the following results given below:

The two figures given below compares the traffic sent/received(packets/sec) in the two scenarios.

The Fig#1 indicates that the average time to send the data packets in a network with a hub or a hub & switch is the same or almost identical.

The Fig#2 indicates that the average time to receive data packets or the throughput in case of a hub & switch network is more than that of a network with only hub.

Fig1. Indicates the time average for traffic sent in packets/sec

Fig2. Indicates the time average for traffic received in packets/sec

The two figures given below compares the collision count and the time delay in the two scenarios.

The Fig#3 indicates that the collision count in a hub and switch network is lesser than in an only hub network.

The Fig#4 indicates that Ethernet delay in a hub and switch network is lesser than in an only hub network.

Fig3. Indicates Ethernet collision count for the two scenarios

Fig 4. Indicates Ethernet delay(sec) for the two scenarios.

The Fig#5, compares the collision count in both the scenarios i.e. with only Hub and in Hub&Switch networks.

It shows that the collision count for only Hub is maximum as compared to the collision count for a Hub&Switch network.

The collision count for the Hub1 and Hub2 in the office network have almost the same collision rate.

Thus, it proves that the collision rate reduces in a network with a switch compared to the network with a hub. Thus, the throughput of switched network is greater than only hub networks.

Fig 5. Indicates the collision count of the Hub in the Only Hub network and the Hub1 and Hub used in the Hub and Switch network.

Question and Answers

Q-1 Explain why adding a switch makes the network perform better in terms of throughput and delay.

Ans: As per the collision count identified for both the network with only Hub and other with a Hub & Switch, the collision count for the Hub1 and Hub 2 in a switched network is much lesser than the Hub1 in the only Hub network. This happens because of the following reasons:

As soon as the Hub receives the packets at its input ports, it forwards them to all outputs irrespective of the destination of the packet to be sent. This increases the chances of collision in hubs.

Where as, in case of a switched network, the switch can receive the packets at its input ports in parallel, and forwards multiple packets to their destined addresses at the same time.

In the lab experiment, there exists two hubs i.e. Hub1 and Hub2 with a switch connecting the two in the middle. Both the hubs receive the packets. The switch acts a a mediator between the two hubs and forwards the packets from one hub to the other with lesser collision.

Thus, the switch helps to reduce the collision rate compared to that with a single hub. Hence, this improves the network performance in terms of throughput and delay.

Q 2 We analyzed the collision counts of the hubs. Can you analyze the collision count of the “Switch”? Explain your answer.

Ans: In the experiment, we have analyzed the collision counts of the hubs in both single and double hubs with a switch networks. From the results obtained from the experiment, it clears tha fact that the collision count using a switch reduces to a greater amount compared to that without a switch. This is due to the fact, that the switch can receive the packets in parallel and buffer the same in case of heavy incoming traffic and also forward the same in parallel to their destined address. Since, there is always a full duplex communication between the switch and the hub, thus the packets exchanged will never collide with each other. Thus, there are hardly any collisions in case of switch.

Q 3 Create two new scenarios. The first one is the same as the OnlyHub scenario but replace the hub with a switch. The second new scenario is the same as the HubAndSwitch scenario but replace both hubs with two switches, remove the old switch, and connect the two switches you just added together with a 10BaseT link. Compare the performance of the four scenarios in terms of delay, throughput, and collision count. Analyze the results. Note: To replace a hub with a switch, right-click on the hub and assign ethernet16_switch to its model attribute.

Ans:

In the first scenario we have duplicated the scenario consisting of only a hub.

Thus, the only hub shown in Fig3.1 has been replaced by a switch as shown in Fig3.2.

In the second scenario, we have duplicated the network with Hub and a switch as shown in Fig#8.

Thus, the two hubs in Fig3.3 have been replaced by two switches and have removed the older switch as shown in Fig3.4. Both the new switches have been connected using a 10 BaseT link.

Fig 3.1 N/W configuration with only hub.

Fig 3.2: N/W configuration with only switch.

Fig 3.3 N/W configuration with two hubs and one switch

Fig3.4 N/W configuration with two switches

Comparing the Results:

The four scenarios which have been compared include the following:

(1) Network with only a hub

(2) Network with a Switch and a Hub

(3) Network with only a switch

(4) Network with two switches.

In the Fig3.5, the graph compares the Ethernet delay in seconds for all the four scenarios. It shows, that the time delay is maximum for a network with a single hub and least for networks with no hubs but switch(s). Thus, the graph shows that the time delay reduces with the number of switches added in the network

The Fig3.6 compares the throughput i.e. the number of packets received per seconds for all the four scenarios. As per the graph, the throughput is almost the same and maximum for the networks consisting of one or two switches. Whereas, the throughput is less with only hub in the network. Thus, the no. of packets received is greater for the networks which are switched based.

Fig 3.5 Compares the Ethernet delay in sec for four scenarios.

Fig3.6 Compares the traffic received (Throughput) in packets/ sec for the four scenarios.

We have not compared the collision count for the four scenarios because the collision count is very less in case of switched networks.

Conclusion

From this lab experiment we have concluded that the switched networks have better throughput, delay and collisions compared to the network with Hubs. The difficulty faced while performing the experiment was the reading of collision count for the switched network with two hubs and the switch in middle. We faced issues in collecting the combined results of collision for the first two scenarios.

But towards the end of the lab experiment, we received good hands on experience on Opnet, and the advantages of having a switched network compared to hubs.