Assignment #4 CS 3357a Fall 2016
Due Dec 2nd, 2016 11:55pm
Background
In Assignment #3 we implemented the rdt 2.2 protocol from the course textbook which allowed us to use UDP for relaible data transfers. Unfortunately, rdt 2.2 only accounted for corrupt packets so we will need to build on our code from Assignment #3 and add a timer to account for lost packets.
Summary
We will once again have two applications (UDP_Client.py and UDP_Server.py) using the UDP protocol to communicate across a network that can corrupt or lose data packets. This means we have to create the reliable data transfer protocol (rdt 3.0) we saw in our textbook by adding a timer to our client application.
The second part of this assignment will be to test your applications using the tc command to create a queing discipline. The specifics are beyond the scope of this assignment but the basic premise is that this command allows us to introduce reordering, delays, and losses in the channel so you can see how your applications perform.
Procedure
Just to refresh your memory, below are the FSMs for both the client and server we need to create (please note – the server/receiver FSM does not change from the last assignment, we still use rdt 2.2):
We will be using the following format for our Pseudo UDP packet:
| ACK | SEQ | DATA | CHKSUM |
ACK – Indicates if packet is ACK or not. Valid values (1 or 0)
SEQ – Sequence number of packet. Valid values (1 or 0)
DATA – Application Data (8 bytes)
CHKSUM – MD5 Checksum of packet (32 Bytes)
You will need to create the packet, load it with the necessary values and then send it to the server. The server will receive the packet, check to see if it is corrupted and then take appropriate action. This process will exactly mirror rdt 3.0 as shown in the textbook, so please make sure you follow it carefully!
Pictured below are four cases that your applications should be able to deal with:
To help you understand the values that would be set for communication in normal, corruption, and data loss situations please reference the chart below:
You must create two files, UDP_Client.py and UDP_Server.py. The details of each are shown below.
UDP_Client –
This app must connect to the UDP_Server app via UDP (you must use the local loopback address of 127.0.0.1 but please choose any port number you wish) then send three separate packets containing the following information:
NCC-1701
NCC-1664
NCC-1017
When adding the timer to the Client.py application, the timeout value should be set to 9ms
Remember, in order to accomplish this, the client application must also be able to receive data in the form of acknowledgements from the server because we will be using the rdt 3.0 process for creating a reliable transfer over UDP.
UDP_Server –
This app will establish a UDP socket and listen for incoming connections from the client. When a connection is made the server will consume the data and follow the rdt 2.2 process as shown in chapter 3 of the course textbook.
Remember, in order to accomplish this, the server application must also be able to send acknowledgements because we will be using the rdt 2.2 process for creating a reliable transfer over UDP.
Configuring Loopback Adapter to Introduce Delays and Losses –
To introduce packet losses and delays we will use this command:
sudo tc qdisc add dev lo root netem delay 10ms reorder 50% 50% loss 40%
This command will delay packets for 10ms (50% of the time) and lose 40% of packets. This queueing discipline runs on your loopback adapter so you must use the loopback address in both your applications.
To revert the loopback adapter to normal behaviour run the following command:
sudo tc qdisc del root dev lo
After you have configured the loopback adapter for delays and losses you will run your applications. Keep in mind that from the screenshots we should be able to see the lost or delayed packets based on what your applications are displaying on the screen.
The output from the UDP_Server and UDP_Client should display a line of text for each of the following actions:
- Received Packet (with all packet values shown)
- Packet Checksum Compare Result (ie. Corrupt or not corrupt)
- Sent Packet (with all packet values shown)
- Timer Expired
Other Information
I have included sample files for both UDP_Client and UDP_Server to get you started. The sample files show you how to do the following:
- Create a UDP connection
- Create a ‘pseudo UDP packet’
- Calculate the checksum
- Fill the packet and send it to the server
- Receive the packet and unpack it
- Compare checksums to ensure the packet is not corrupted
You will need to do the rest.