CSC3002 – Networks Assignment 1 – 2024

Socket programming project

Department of Computer Science
University of Cape Town, South Africa
February 19, 2024

This assignment is on networked applications where you will learn the

basics of protocol design and socket programming with TCP and UDP con-
nections in Python: how to create a socket, bind it to a specific address
and port, as well as send and receive messages/files. You will develop a
client-server + p2p application in groups of three students. This document
describes the context/requirements (Section 1), what to submit (Section 3),
and some basic information about socket programming (Section A).

1. Application Description
Chat messaging applications require an instant messaging protocol that allows for (near)real-
time communication between clients. Both sender and receiver must follow the same ap-
plication protocol. Some of the well known protocols include WebRTC, WebSocket, and
IRC - Internet Relay Chat. Most of these protocols work by establishing peer-to-peer
connections between client applications; this means messages are sent directly between
two devices without going through a server. However, in some cases, a server may still
be involved for tasks such as signaling and initialising/facilitating the establishment of
the peer-to-peer connections.
In this assignment, you are to design and implement a peer-to-per chat application
that makes use of both UDP and TCP sockets – UDP for transmitting real-time media
streams; TCP for signaling and initializing peer communication. A server is needed to
coordinate initial client connections, and clients should be able to query the server for
lists of available clients, and to obtain parameters for the establishment of communication
between clients.
You should implement your client to use a TCP connection to interact with server.
The client implementation should allow for command line arguments, such as specifying
the server IP address, the port at which the server is listening, and the request, e.g to
sign-up or announce connection parameters. Clients may also want to request a list of

1
available peers, or communication parameters of a particular peer. Servers needs to have
means for maintaining a list of active peer clients.

2. Protocol Design and Specification

An important step in protocol design involves defining the framework of communication
for the intended application. This entails specifying requirements and constraints such
as: whether real-time interaction is expected; reliability (ie if we need to verify/check that
every message is delivered correctly); and authentication/confidentiality issues, among
others. In this assignment, you are to design the application that takes into consideration
privacy/confidentiality, by allowing users indicate information visibility permissions, e.g
does a client want to be visible to anyone querying the server? Or, should a file be
visible (or not) to certain specific clients? Or should a shared secret-key be used to
prove authenticity of the clients/messages? The application protocol should be designed
to include some of these features.
Specification of protocol messages involves defining the types and structure of mes-
sages. Three types of messages can be defined; commands, data transfer, and control.
Command messages define the different stages of communication between parties, such
as the initiation or termination of communication. Data transfer messages are used to
carry the data that is exchanged between parties, and such data could be fragmented
into several messages. Control messages manage the dialogue between parties, including
such aspects as message acknowledgements, and re-transmission requests.
The message structure constitutes at least the header and body. The header, whose
structure must be known to both sides of a communication link, may contain fields that
describe the actual data in the message. Some of the fields/information contained in the
header might be the message type, the command, recipient information, and sequence
information. The header generally has fixed size and contains clues that should help the
receiver to understand the rest of the message.
The last aspect of the protocol design will be the communication rules that specify
the sequence of messages at every stage of communication. For example, a peer can be
in ’available’, ’connected’, or ’away’ state. How do the peers respond to messages while
in these different states, and how do they transition between states. You need to specify
messages and reactions for different states. You will need to represent such rules with
sequence diagrams.

3. What you need to submit

As a group, design/specify a protocol to support the application, and implement the
server. As individuals within each group, implement a client that can interact with the
group’s server and other clients. You will be required to submit the following:
1. The group server code, and individual client code, with proper inline documenta-
tion (comments)

2
2. A report (max 6 pages) on the design and functionality of your chat application
(individual sections on the client implementation).

3. In addition, the report needs to include:

a) A list of features with a brief explanation for their inclusion
b) A protocol specification, detailing the message formats and structure. You
are required to include sequence diagram(s).
c) Screenshots of the application revealing its features.

4. Oral presentation to be scheduled with the TAs and Tutors (oral to be done on
day after submission deadline)

3
A. Multi-threaded Client/Server Applications—Sockets
Programming
A.1. What is a socket?
A socket is the one end-point of a two-way communication link between two programs
running over the network. Running over the network means that the programs run on
different computers, usually referred as the local and the remote computers. However
one can run the two programs on the same computer. Such communicating programs
constitutes a client/server application. The server implements a dedicated logic, called
service. The clients connect to the server to get served, for example, to obtain some
data or to ask for the computation of some data. Different client/server applications
implement different kind of services.
To distinguish different services, a numbering convention was proposed. This con-
vention uses integer numbers, called port numbers, to denote the services. A server
implementing a service assigns a specific port number to the entry point of the service.
There are no specific physical entry points for the services in a computer. The port num-
bers for services are stored in configuration files and are used by the computer software
to create network connections.
A socked is a complex data structure that contains an internet address and a port
number. A socket, however, is referenced by its descriptor, like a file which is referenced
by a file descriptor. That is why, the sockets are accessed via an application programming
interface (API) similar to the file input/output API. This makes the programming of
network applications very simple. The two-way communication link between the two
programs running on different computers is done by reading from and writing to the
sockets created on these computers. The data read from a socked is the data wrote
into the other socket of the link. And reciprocally, the the data wrote into a socket
in the data read from the other socket of the link. These two sockets are created and
linked during the connection creation phase. The link between two sockets is like a
pipe that is implemented using a stack of protocols. This linking of the sockets involves
that internally a socket has a much more complex data structure, or more precisely, a
collaboration of data structures. Thus, a socket data structure is more than just an
internet address and a port number. You have to imagine a socket as a data structure
that contains at least the internet address and the port number on the local computer,
and the internet address and the port number on the remote computer.

A.2. How is a network connection created?

A network connection is initiated by a client program when it creates a socket for the
communication with the server. To create the socket, the client calls the Socket con-
structor and passes the server address and the the specific server port number to it. At
this stage the server must be started on the machine having the specified address and
listening for connections on its specific port number.
The server uses a specific port dedicated only to listening for connection requests from

4
clients. It can not use this specific port for data communication with the clients because
the server must be able to accept the client connection at any instant. So, its specific
port is dedicated only to listening for new connection requests. The server side socket
associated with specific port is called server socket. When a connection request arrives
on this socket from the client side, the client and the server establish a connection. This
connection is established as follows:

1. When the server receives a connection request on its specific server port, it creates
a new socket for it and binds a port number to it.

2. It sends the new port number to the client to inform it that the connection is
established.

3. The server goes on now by listening on two ports:

• it waits for new incoming connection requests on its specific port, and
• it reads and writes messages on established connection (on new port) with
the accepted client.

The server communicates with the client by reading from and writing to the new port.
If other connection requests arrive, the server accepts them in the similar way creating
a new port for each new connection. Thus, at any instant, the server must be able
to communicate simultaneously with many clients and to wait on the same time for
incoming requests on its specific server port. The communication with each client is
done via the sockets created for each communication.
TCP is a connection-oriented protocol. In order to communicate over the TCP proto-
col, a connection must first be established between two sockets. While one of the sockets
listens for a connection request (server), the other asks for a connection (client). Once
the two sockets are connected, they can be used to transmit and/or to receive data.
When we say ”two sockets are connected” we mean the fact that the server accepted a
connection. As it was explained above the server creates a new local socket for the new
connection. The process of the new local socket creation, however, is transparent for the
client.
The datagram communication protocol, known as UDP (user datagram protocol), is a
connectionless protocol. No connection is established before sending the data. The data
are sent in a packet called datagram. The datagram is sent like a request for establishing
a connection. However, the datagram contains not only the addresses, it contains the
user data also. Once it arrives to the destination the user data are read by the remote
application and no connection is established. This protocol requires that each time a
datagram is sent, the local socket and the remote socket addresses must also be sent in
the datagram. These addresses are sent in each datagram.