The Basics Of Designing A Network Information Technology Essay

Objective of this lab is to demonstrate the basics of designing a network, taking into consideration the users, services and locations of the hosts.

Overview:

The major purpose is designing a network and optimizing it to have a maximized network performance, taking into consideration the cost constraint and the required services to be offered to different types of users. Periodical network optimization should be done throughout the whole life of network to get maximum network performance as well as to monitor the utilization of network resources.

In this lab we will design a network for a company having four departments which includes Engineering, E-Commerce, Research and Sales. We will utilize a LAN model that allows simulating multiple clients and servers in one simulation object. This will reduce the amount of configuration work as well as amount of memory required to execute the simulation. After completing this laboratory we will be able to study how different design affects the performance of the network.

Implementation:

Scenario 1 (Simple Network):

We will design a campus network

We will following objects in the network:

Application Config: It will be used in configuration of profiles for users.

Profile Config: It will describe profile for users.

Subnet

We will configure subnets as Engineering, Research, E-Commerce and Sales.

Each node has a 10BaseT LAN and Ethernet_16Switch connected by 10BaseT link.

We will create 4 profiles with respect to the nodes which are Engineer, Researcher, E- Commerce Customer and Sales Person.

The main server subnet has one Ethernet_16switch and three Ethernet_servers connected by 10BaseT link. These servers are web server, file server and database server.

Web server supports the services like Web Browsing (Light HTTP1.1), Web Browsing(Heavy HTTP1.1), E-mail (Light) and Telnet Session (Light)

File Server supports the services like File Transfer (Light) and File Print (Light).

Database Server provides the service of Database Access (Light).

100BaseT link is used to connect all the four subnets with the server subnet. Thus there will be a connection between profile switch as a node A and server switch as node B.

The final network will look like Figure 1.

Figure 1: Network Design

Statistics: We used Page Response Time from the HTTP protocol for the testing the networks performance.

Simulation: Duration for simulation will be set to 30 minutes.

Duplicate Scenario (Busy Network):

In this scenario we will set the Background Utilization between the 2 servers to 99% instead of 0%.

Result:

After simulating the both scenarios we will get result for Page Response Time as shown in Figure 2.

Figure 2: Graph for Page Response Time for both scenarios.

It can be seen from the Figure 2 that Page Response Time for busy network is always higher than the simple network. It also being noticed that in the beginning page response time increases for both the networks but after some time they will be in stable state.

Analysis:

It can be seen that page response time for busy network is higher than the simple network. This is because we set the Background Utilization to 99% in Busy Network. Thus as the CPU Utilization increases the page response time decreases.

Expectation:

As we have set the Background utilization to 99% in the Busy network, we are expecting higher page response time in this scenario. Result of the laboratory is identical to our expectation.

Questions / Answers:

Analyze the result we obtained regarding HTTP page response time. Collect four other statistics of your choice and rerun the simulation of the Simple and the Busy network scenarios. Get the graphs that compare the collected statistics. Comment on these results.

Page Response time: Page response time is higher in Busy network than simple network. This higher page response time is because of background utilization is 99%.

Figure 2: Graph comparing Page Response Time for both scenarios.

Object Response Time: It can be seen in Figure 3 that Object Response Time is very less for Simple network in compare to Busy network. Less response time means that access of the object in the network is higher.

Figure 3: Graph comparing Object Response Time for both scenarios.

Upload Response Time: It can be seen in Figure 4 that Upload Response Time for Simple network is constant throughout the simulation while for the Busy network it will vary. As network is busy it takes more time uploading in FTP connection for Busy Network while in Simple network uploading time is less.

Figure 4: Graph comparing Upload Response Time for both scenarios.

Download Response Time: Downloading a file in a busy network is high compare to simple network. It can be also seen from Figure 5 that downloading time for simple network is constant while for busy network it is varying with time.

Figure 5: Graph comparing Download Response Time for both scenarios.

Ethernet Delay: It can be seen in Figure 6 that delay for busy network is higher than simple networks. This proves that delay is directly proportional to the background utilization.

Figure 6: Graph comparing Ethernet Delay for both scenarios.

In the BusyNetwork scenario, study the utilization% of the CPUs in the servers.

CPU Utilization:

Web Server: CPU utilization increases in the early state and then become constant for the remaining period of time.

Figure 7: CPU Utilization for Web Server in BusyNetwork .

File Server: CPU utilization goes high as we access the File Server but letter it will decreases exponentially.

Figure 8: CPU Utilization for File Server in BusyNetwork.

Database Server: CPU utilization will go high as database server is accessed and it will remain almost constant with very less fluctuation for remaining time span.

Figure 9: CPU Utilization for Databse Server in BusyNetwork.

Create a new scenario as a duplicate of the BusyNetwork scenario. Name the new scenario Q3_OneServer. Replace the three servers with only one server that supports all required services. Study the utilization% of that server’s CPU. Compare this utilization with the three CPU utilization you obtained in the previous question.

One Server:

Combining all the server in one server doesn’t make much difference in CPU utilization. CPU utilization for BusyNetwork and Q3_OneServer is very similar.

The utilization of one server is very much similar to the utilization for databse server and web server. It is not matching with the File server utilization as one server is serving all three purposes it is not decreasing unlike in File server.

Figure 10: CPU Utilization for Server in OneServer Network.

Create a new scenario as a duplicate of the BusyNetwork scenario. Name the new scenario Q4_FasterNetwork. In the Q4_FasterNetwork scenario, replace all 100baseT link with 10Gbps Ethernet links and replace all 10BaseT links with 100BaseT links. Study how increasing the bandwidth of the link affects the performance of the network in the new scenario (e.g. compare the HTTP page response time in the new scenario with that of the BusyNetwork).

In this new scenario we are increasing the speed of the link which will directly increase the speed of network. Faster network have less response time as data rate is high compare to BusyNetwork. Figure 11 shows that pager response time is less for FasterNetwork is less in compare to BusyNetwork because of the high data rate.

Figure 11: Graph for page response time of Busy and Faster Network

Conclusion:

Through this laboratory we learned to design a network, taking consideration of the users, services and locations of the hosts. It has also demonstrated the effect of devices on CPU utilization.