Socket programming using C

CN LAB 2016
1
1
Introduction to Socket
Programming
Part-I
Ajit K Nayak, Ph.D.
2
Sockets!!!
• Is it like this ???
• Electrical Sockets, used to transfer
Electrical power

CN LAB 2016
2
3
Sockets in Network Programs
Server
Socket
Client
Socket
Network
010011001001111
Client Program
Server Program
Consumed in
Application Layer
010011001001111010011001001111
4
Network Sockets
• The socket is the method for accomplishing
communication among processes in a network.
• An interface between application and the network
• i.e. The application can send/receive data to/from the
network via sockets
• In UNIX terminology you can say that it is also a file,
– As a file is understood by a file descriptor
– Socket is also recognized by a socket descriptor ( a 16 bit
integer)
• Sockets are always created in pairs and they are used to
communicate with other sockets
• There are many kinds of sockets like Internet sockets,
UNIX sockets, . . .

CN LAB 2016
3
5
Socket API
• We will use Socket API with C programming in
Linux environment
• Socket API contains many data structures,
functions and system calls which will help us to
write good and efficient network programs
• Requirement is any PC loaded with Linux, C and
Socket API
• It is also possible to program in other
environments like windows and other UNIX
systems!!! You need to explore
6
• SOCK_STREAM
– TCP
– reliable delivery
– in-order guaranteed
– connection-oriented
– bidirectional
• SOCK_DGRAM
– UDP
– unreliable delivery
– no order guarantee
– no notion of “connection”
– can send or receive
App
socket
3 2 1
Dest.
App
socket
3 2 1 Dest
Essential types of Sockets

CN LAB 2016
4
7
Stream Oriented Comm.
• Create Sockets at both sides
• Bind it to a local port in server
side
• Listen for a incoming
connection request (server)
• Connect from a client
• Accept the request (server)
• Talk to each other
• Close the sockets
• Also called TCP sockets
Server Client
socket
send
bind
listen
recv
Connect
close
send
accept
recv
close
.
.
.
socket
8
Datagram Socket
• Create Sockets
• Bind Server Socket
•Talk to each other
• Close the socket
• Also called UDP sockets
Server Client
socket
sendto recvfrom
close
sendtorecvfrom
close
.
.
.
socket
bind

CN LAB 2016
5
9
Data Structures
struct in_addr{
in_addr_t s_addr;
};
• 32 bit unsigned net/host id in network byte order
• typedef uint32_t in_addr_t;
• Network byte order is in big-endian format and
host byte order is in little-endian format
10
IPV4 Socket Add Structure
struct sockaddr_in{
uint8_t sin_len;
sa_family_t sin_family;
in_port_t sin_port;
struct in_addr sin_addr;
char sin_zero[8];
};
• sin_len: length of structure
• sin_family: socket address family
• sin_port_t: 16 bit unsigned integer
• sin_port: port no of TCP or UDP
• sin_addr: IP address

CN LAB 2016
6
11
System Calls
 Creating a socket
int socket( int family, int type, int protocol);
• On success the socket function returns a non-negative integer,
called socket descriptor
• This call is used by both server and client
• sockfd = socket(AF_INET, SOCK_STREAM,0);
• family specifies the protocol family
Family Description
AF_INET IPv4 protocols
AF_INET6 IPv6 protocols
. . .
12
System Calls(2)
– type: communication type
type Description
SOCK_STREAM stream socket
SOCK_DGRAM datagram socket
SOCK_RAW raw socket
– protocol: specifies protocol, usually set to 0
except for raw sockets (so that socket choose
correct protocol based on type)

CN LAB 2016
7
13
System Calls(3)
int bind(int sockfd, const struct sockaddr *myaddr,
socklen_t addrlen);
• ‘bind’ assigns a local protocol address to a
socket
• Local protocol address consists of IP address
along with a port number
• Second argument is a pointer(?) to address
structure
• Third argument is the length(size) of address
structure (32-bit integer)
14
System Calls(4)
serv_addr.sin_addr.s_addr=htons(INADDR_ANY)/* wild card */
serv_addr.sin_port = 0;
bind(sockfd, (struct sockaddr *)&serv_addr,
sizeof(serv_addr);
• It returns ‘0’ if OK or ‘–1’ on error
• Normally a TCP client does not bind an IP address to its
socket. The Kernel chooses the source IP when socket is
connected, based on outgoing interface.

CN LAB 2016
8
15
• With TCP, calling bind lets us specify a port
number, an IP address, both or neither.
Process specifies Result
IP address Port
wildcard 0 Kernel chooses IP address and
port
wildcard nonzero Kernel chooses IP address,
process specifies port
Local IP 0 Process specifies IP address,
Kernel chooses port
Local IP nonzero Process specifies IP address,
and port
System Calls(5)
16
Byte ordering functions
#include <netinet/in.h>
 uint32_t htonl(uint32_t hostlong);
 it converts the long integer hostlong from host byte order to network byte
order.
 uint16_t htons(uint16_t hostshort);
 it converts the short integer hostshort from host byte order to network byte
order.
 Both returns value in network byte order
 uint32_t ntohl(uint32_t netlong);
 it converts the long integer netlong from network byte order to host byte
order.
 uint16_t ntohs(uint16_t netshort);
 it converts the short integer netshort from network byte order to host byte
order. Both returns values in host byte order

CN LAB 2016
9
17
System Calls(6)
int listen(int sockfd, int backlog);
• The listen function converts an unconnected
socket into a passive socket, indicating that the
kernel should accept incoming connection
requests directed to this socket
• The call to listen moves the socket from the
CLOSED state to LISTEN state
• The second argument specifies the maximum
number of connection that the kernel should
queue for this socket
listen(sockfd,5);
18
System Calls(7)
int connect(int sockfd, const struct sockaddr
*servaddr, socklen_t addrlen);
• Used by a client to connect to a server.
• The connect function initiates TCP’s three-way
handshake process.
• The function returns only when the
connection is established(0) or when an error
occurs(-1)
• If the connect fails, the socket is no longer
usable and must be closed.

CN LAB 2016
10
19
TCP Connection Establishment
Socket
connect(blocks)
(active open)
connect returns
Client Server
Socket, bind,
listen, accept
(blocks)
1. The server is in
passive open state
4. The client acknowledges the server’s SYN
The minimum no of packets required for this exchange is three; hence is
called three-way handshake
3. Server acknowledges client’s
SYN sends its own SYN
2. The client issues an active
open by calling connect. I.e
it sends a SYN segment
20
System Calls(8)
int accept(int sockfd, struct sockaddr *cliaddr,
socklen_t *addrlen);
• It is called by TCP server to return the next completed
connection
• The cliaddr and addrlen arguments are used to return
the protocol address of the connected peer process.
• The addrlen contains the sizeof client address structure
and on return it contains the actual no of bytes stored by
the kernel in the socket address structure

CN LAB 2016
11
21
System Calls(9)
• If accept is successful, it returns a new descriptor
(socket) that was automatically created by the
kernel.
• This new descriptor refers to the TCP connection
with the client for data communication
• Now first one is called the listening socket (sockfd)
and the second one is called the connected socket
(connfd)
• A given server normally creates only one listening
socket, which exists for the life time of the server.
• The kernel creates one connected socket for each
client connection that is accepted.
22
System Calls(10)
addrlen = sizeof(cliaddr);
connfd = accept(sockfd, (struct sockaddr *)
&cliaddr, &addrlen);
• This function gives up to three values
– An integer return, that is either a new socket
descriptor or an error indication(-1)
– The protocol address of the client process
(through cliaddr) and
– Size of this address (through addrlen)
• If protocol address is not required then both
cliaddr and addrlen is set to NULL;

CN LAB 2016
12
23
System Calls(11)
int close(int sockfd);
• The default action is to mark the socket as closed
and return to process
• The socket descriptor is no longer usable by the
process; It cannot be used to read or write further
• But TCP will try to send any data that is already
queued to be sent to other end, and after this the
normal TCP connection termination sequence takes
place.
24
Network I/O- Reading
• Connection oriented
int recv(int sockfd, char *buf, int nbytes, int tag);
– It reads nbytes(max) from socket to the buffer
– Returns no of bytes read successfully from the socket and
–1 on error
• Connection less
int recvfrom(int sockfd, char *buf, int nbytes, int flag,
struct sockaddr *from, int addrlen);
– It receives nbytes(max) from a socket, whose address is
given by the from address structure to the buffer
– Returns no of bytes read successfully from the socket and
–1 on error

CN LAB 2016
13
25
Network I/O- Writing
• Connection oriented
int send(int sockfd, char *buf, int nbytes, int tag);
– It writes n bytes(max) to the socket from buffer
– Returns no of bytes written successfully to the socket and
–1 on error
• Connection less
int sendto(int sockfd, char *buf, int nbytes, int flag,
struct sockaddr *to, int addrlen);
– It sends nbytes(max) to a socket, whose address is given by
the to address structure from the buffer
– Returns no of bytes sent successfully from the socket and –
1 on error
26
References
• W. Richard Stevens, UNIX network
programming, vol 1, PE
• Beej’s Guide to Network Programming using
internet sockets
http://beej.us/guide/bgnet/
• M. J. Donahoo, K. L. Calvert – Pocket Guide to
TCP/IP Sockets, Harcourt Indian 2001.

Network Programming/BKM & AKN/Part II
Unix Network Programming
Part II
Prepared byPrepared by
Bimal K Meher & Ajit K Nayak
Department of CSEA, SILICON

Concurrent Server
• Concurrent server is a server, which can handle multiple
clients simultaneously. i.e. Multiple clients can talk to a
single server at a time.
• The servers we have designed so far are called iterative
servers. i.e. One client talks to the server at a time, another
client can talk if the previous client closes the connection
• The examples is: many telnet clients from our lab (each
user has a client) is connected to one telnet server and all
can simultaneously execute their commands.
• To design a concurrent server the fork system call of unix
is used

fork() System Call
• fork() is a UNIX function which can create a duplicate of a process called
child process.
• fork() takes no arguments, returns integer
• fork() called once, it returns twice
– Once in the calling process called parent process, with an integer value called
as ppid (parent process ID)
– Once in the newly created process called child process, with a value 0
• The child may call getppid() to get parent process ID
• Child is an exact copy of parent process
• All descriptors opened in the parent is shared by child and the program
execution resumes from the statement after fork

Using fork(), an example
• /* Demo of simple use of fork() */
int main() {
fork();
printf(“nHello How r u ?n”);
}
• Output: Hello How r u ?
Hello How r u ?
• Assignments
– Now try using 2 forks instead of 1.
– Print the ppid of each process
– Print the pid of each process
– Print different messages in parent and child

getpid() vs getppid()
• Now identify the missing links & tell the output.
• int main(){
int pid;
pid=fork();
if (pid==0) {
prinf(“nI am the my PID is %d ”, getpid());
}
else
printf(“nI am the my PID is %d” , getpid());
}
child
parent
printf(“nMy parent’s PID is %d”,getppid());
• By using fork(), a process makes a copy of itself so that one copy(say
parent) can handle one operation while the other copy(say child) does
another task.

An Iterative Server
connect()
Client1
listenfd
Step 1: Diagram shows the status of the client & server while the
server is blocked in the call to accept & the connection re
quest arrive from the client.
Step 2: After accept returns, the connection is accepted by the
kernel & a new socket i.e. connfd is created.
sockfd listenfd
connfd

An Iterative Server
Step 3: when first connection closes, the queued request is accepted
with a new connfd
Step 3: Suppose another client requests, then it will be queued till
the first connection closes
Client1
connect()Client2 listenfd
sockfd
connfd
connect()Client2
listenfd
sockfdClient1
closed
connfd

Designing Concurrent Server
• The idea is once a child process is created, then that
process can handle the incoming request from the
client.
• Depending on how many clients are there in the
system that many child can be forked by the server
process.
• In our case when we use a fork after accept(), it
creates a child process out of server process.
• So that the child can handle the new connection from
the client & parent remains alert for another new
connection request from another client.
• This is the basis of Network Servers.

Concurrent Server
connect()
Client1 Server
listenfd
Step 1: Diagram shows the status of the client & server while the
server is blocked in the call to accept & the connection re
quest arrive from the client.
Step 2: After accept returns, the connection is accepted by the
kernel & a new socket i.e. connfd is created.
sockfd
listenfd
connfd

Status of client/server after fork returns
Step 3: When fork() is called, a child is created. So both the
descriptors i.e.listenfd & connfd are shared(duplicated)
between the parent & child.
sockfd
listenfd
listenfd
connfd
connfd
Client1 Server(parent)
Server(child)

Status of client/server after close()
Step 4: The next step is, the parent closes the connected socket
& the child closes the listening socket.
connect()
listenfd
connfd
Client1 Server(parent)
Server(child)
connect()
Client2
Note:
This is the desired final state of the sockets. The child is handling the
connection with the client & the parent can call accept() again on the
listening socket to handle the next client connection.

Outline of a typical Concurrent Server
listenfd=socket(….);
……/* other source codes */
bind(listenfd,…..);
listen(listenfd,qlen);
for ( ; ; ) {
connfd=accept(listenfd,…..);
if ((pid=fork())= =0){ /* true for child process */
close(listenfd) /* child closes listening socket */
/* process the request with help of connfd */
close(connfd); /* done with this client */
exit(0); /* child terminates */
}
close(connfd); /* parent process */
}

Thank You

Socket programming using C

Recommended

More Related Content

What's hot (20)

Similar to Socket programming using C (20)

More from Ajit Nayak (20)

Recently uploaded (20)

Socket programming using C