5NTCM004W Network Simulation Coursework 2 Analysis and Results

This document presents a comprehensive solution to a network simulation coursework, addressing the assignment's requirements. The solution begins with the setup of Riverbed Modeler Academic Edition, detailing the creation of a project and scenario. It then focuses on collecting and analyzing FTP and email traffic statistics in both simple and busy network scenarios, comparing the results graphically. The assignment further explores statistical analysis, including calculating confidence intervals and discussing the advantages and disadvantages of data replication. Finally, it addresses the calculation of confidence intervals for average wait times and determining the necessary number of replicas for optimal performance. The solution offers insights into network design, simulation techniques, and the practical application of statistical concepts in network analysis.

Name:
Professor:
Course:
Date:

a.
b.
i. Start Riverbed Modeler Academic Edition and Choose New from the File menu.
ii. Select Project and click OK, Name the project and the scenario you are designing and Make
sure that the Use Startup Wizard and is checked then Click OK.
iii. In the Startup Wizard: Create Empty Scenario, Click Next, Choose Miles from the Size
drop-down menu and assign 1 for both X Span and Y Span, Click Next twice, Click Finish.
a. Collect the following two statistics:
i. FTP Traffic received (packet/sec)
ii. E-mail Traffic Received (packet/sec)
Rerun the simulation of the Simple and the Busy network scenarios. Obtain the graphs that
compare the collected statistics and comment on these results

In both FTP and email scenarios Simple network a page takes less time to respond more time is
taken for response in busy networks
d)
i.
For 95% confidential interval, the z* value will be 1.96
Example
Replication x
1 201.34
2 205.64
3 203.45
4 207.85
5 204.23
6 209.76
7 201.45
8 204.896
9 243.76
10 563.89
Mean 2446.27/10=244.6266
Thus For 95% confidence level,
Mean*1.96/sqrt of 10
=244.6266*1.96=479.4681
479.4681/3.1623
=77.36

ii.
Advantages
Crashes problems are minimized. If the system has many replicas, the it means that if one them crashes
at specific time the the other replica will be in use. One replica in the system a can be used to substitute
another replica in the same system and thus there will be protection against corrupted data in the system.
Due to increase in the count of operations sets that are needed to gain access to the data which is being
controlled by the server, the overall performance and efficiency is improved through replication of the
server and the work is also divided.
Through independent replication, the replicated data will be available to the client even if the local
copy of the data is not available. This means that all the replicated data will be available to the client at all
sites in the network.
Drawbacks
It is potentially expensive when the replicas at all different sites need to be updated.
It is difficult to maintain data consistency. It requires complex measures to attain the consistency of
data when independent replication is used. Through independent replications, the original data will have
many replicas. At such a case, when modification is done on one of the replica in the site, then the
changes will be quite different from the rest of the replicas. If at a certain case there are changes that are
being done to one of the copy and the same is not duplicated to other copies, this will lead to other copies
being out of date thus causing conflicts.
A lot of space is required for storing the replicas which belongs to the same original data at different
sites thus causing the memory problems.
It will require the cost of the bandwidth to be increase in order to maintain the independent
replication.
Independent replications require the data in files to be updated. In a network the user gains access and
interacts with the data in files and he/she will at times modify or delete this data. As a result, large
number of the messages count will be flowing through the network causing independent replications to be
expensive.
e)
i.
General Average wait =1.56+ 1.10+ 1.89+ 1.24+ 0.99+ 1.81 +1.59+ 1.02 +1.56 +1.24
=14/10
=1.4
90% confidence interval=1.96
1.4 * 1.96=2.744
= 2.744
√n

=0.8677
Interval=2.3077, 2.2677
ii.
90% confidence interval=1.96
1.4 * 1.96=2.744
= 2.744
√n
=0.8677
Resulting interval=0.8677*0.15=1.3016
Replicas count=1.3016*10=1.302
Two additional replications
iii.
The replicated data will be available to the client even if the local copy of the data is not available.