EE2702代写、Programming代做、代做Python语言、Python编程调试

日期：2020-04-18 11:24

EE2702 Computer Science and Programming

2019/2020

Coursework Assignment 2 – Discrete Event Simulation of an Internet Router

General information:

You need to submit on Moodle a single .py document which should ideally be titled yourname.py

before the deadline of 20 April 2020.

The submitted code should perform the functions of the assignment presented in this document.

Assignment:

Simulation is a method of analysing real-world events or newly designed algorithms and solutions. The

discrete event simulation model consists of identifying individual events in the simulated model and

assigning these events a unique time in the time schedule. Discrete event simulation can be used to

simulate a number of real systems, including computers networks, biomedical processes, supply

chains, and a variety of scheduling problems.

For this assignment, you are expected to produce a python code which performs a full discrete event

simulation of a very simple Internet router. The router is characterized by:

- Random arrival of packets

- Random distribution of packet sizes

- Multiple router output interfaces (‘servers’)

- Serving schedule – the simulation will have to support both first-in-first-out (FIFO) serving,

and also priority serving

The case study used for this assignment is operation of a very simple Internet router. Internet packets

of variable sizes arrive at the router to be processed and forwarded further in the networks. Your

simulation will enable performance evaluation of the router, by identifying the average waiting time,

the probability of immediate response, the average packet size, and other performance parameters.

Assumption is made that the router has fixed output data rate – the duration of the packet processing

can be considered identical to the packet size.

To develop the simulator, you are asked to complete the following challenges:

2.1 – one server, FIFO queueing

In this first assignment, you are asked to develop a discrete-event simulation of an Internet router

with a single input queue and a single output interface. The packets arrive following a random Poisson

process, with a pre-defined average time between two arriving packets. The Poisson process is

characterized with an exponential distribution of inter-arrival times. The packets have random sizes –

the packet size should be calculated by the simulator at the packet arrival. The packet size should be

considered a random variable with exponential distribution. In the case the packet has arrived and

cannot be immediately served, it has to wait for the first available opportunity to be served. The packet

size determines the service time duration.

Your code should request the average system parameters – average packet size and average interarrival

time for the packets, and should calculate the following parameters:

• average waiting time

• average queue size

• probability that the queue size is 0 at arrival of a new call

• probability that the queue size is greater than 5 times the average packet size.

For the presentation of the simulation results, see section ‘Presenting Simulation Results’ below.

2.2 – multiple servers, FIFO queueing, random choice of server

In the second assignment, your code from the previous assignment should be modified to support

four separate servers (four output router interfaces). In the simulation, upon the arrival of the packet,

the server (output interface) should be randomly identified for that packet, regardless of the queue

sizes.

Your code should request the average system parameters – average packet size and average interarrival

time for the packets, and should calculate the same simulation results as in assignment 2.1.

Similarly to 2.1, generate graphs to show how calculated parameters change with the change in the

input values – the average packet size or the average packet inter-arrival time.

For this assignment only, make sure the graphs show the simulated results for each server.

2.3 – two servers, two classes, random assignment of classes with mean probabilities 20-80, one

queue for the priority class and one for the non-priority

In the third assignment, your simulator should operate as in the previous assignments, but this time

at the generation (“arrival”) of the packets the packets should be assigned a class (“priority” or

“economy”). The class assignment should be random, but on average 20% of the packets should be

“priority” and 80% should be “economy”. The serving policies should be as follows:

• Policy A: one server should be dedicated for the priority packets only, and one for the

economy packets. If there are no priority packets, that server should be empty

• Policy B: one server should be dedicated for the priority packets, and one for the economy

packets. If there are no priority packets, the first waiting economy packet should be served.

Your code should request the average system parameters – average packet size and average interarrival

time for the packets, probability of the priority packets (default = 20%) and should calculate

the following parameters:

• average waiting time for both traffic classes

• average queue size for both traffic classes

• probability that the queue size for each of the classes is more than 5

• probability that the queue size is 0 at arrival of a new packet

Again, your program should generate graphs to show how calculated parameters change with the

changes in the input values.

What conclusion can you make when you compare the performance of policy A and policy B?

Presenting Simulation Results

For all simulations:

• The mean values for these parameters should be calculated in the simulation and printed on

the screen.

• In addition to this, the simulation(s) should generate a time graph, to show how average

values for the average waiting time and the average queue size change with simulation time.

• The simulation should also show how the parameters given above change with the increase

in the average packet size or with the increase in the average inter-arrival time. This is best

done using graphs, with values for average packet size / average inter-arrival time on x-axis,

and calculated parameters on the y-axis.

• When running simulations, make sure the simulation run to process at least 10,000 packets.

Assessment and Marking scheme

Following the submission of your simulation code, one-to-one examinations will take place in Dr

Rakocevic’s office C130. The schedule for this will be available later.

Each examination will last about 15 minutes. You will be asked to run your simulation, and to explain

the operation of the code.

Please make sure the user interface for your simulation is ready and easy to use. If you wish to use

your own computer, make sure it is connected to the University network, and all the libraries are

installed.

Your code will be carefully examined even in the case that the code is not completely functional. A

very rough distribution of maximum marks is:

- full operation of the code and presentation of all simulation results solving challenge 2.1 :

maximum mark of 65%

- full operation of the code and presentation of all simulation results solving challenges 2.1 and

2.2: maximum mark of 85%

- full operation of the code and presentation of all simulation results solving challenges 2.1, 2.2,

and 2.3: maximum mark of 100%