UDP Echo Server In C Tutorial Part II

The second part to the UDP Echo Server in C tutorial. In part II we add the server component of the code base and together with the client code from part I we’ll have a functioning application.

Intro

Click here for Part I

Welcome to the second part of the how to write a echo server in C. In this short second part we’ll finish up by adding the server code that our client in part I will use connect to over UDP/IP. Once again the code will work on all the major platforms Mac, Linux and Windows. Much of the socket address setup code and creation of a file descriptor is the same as the client application but the major difference with the server is that the server must bind its socket to port.

• Move into the working directory created in Part I called UdpEcho
• Create a file called server.c

Part II: The Server

If you read through Part I of this tutorial much of the code here will be instantly familiar. The server code uses the same platform object macros and selects a branch of code based on the system its being executed on.

C preprocessor setup, function prototypes and main

/* server.c */
#define PLATFORM_WINDOWS  1
#define PLATFORM_MAC      2
#define PLATFORM_LINUX    3

#include <stdio.h>
#include <stdlib.h>
#include <string.h>

#if defined(_WIN32)

#define PLATFORM PLATFORM_WINDOWS

#elif defined(__APPLE__)

#define PLATFORM PLATFORM_MAC

#else

#define PLATFORM PLATFORM_LINUX

#endif

#if PLATFORM == PLATFORM_WINDOWS

#include <winsock2.h>

#elif PLATFORM == PLATFORM_MAC || PLATFORM == PLATFORM_LINUX

#include <sys/socket.h>
#include <netinet/in.h>
#include <fcntl.h>
#include <arpa/inet.h>

#endif

#if PLATFORM == PLATFORM_WINDOWS

#pragma comment(lib,"ws2_32.lib")

#endif

#define PORTNUM 12354
#define BUFMAX 1024

#if PLATFORM == PLATFORM_MAC || PLATFORM == PLATFORM_LINUX
  void macLinuxEchoLoop(int, struct sockaddr*, socklen_t);
  void macLinuxEchoServer();
#endif

#if PLATFORM == PLATFORM_WINDOWS
  void windowsEchoLoop(SOCKET, sockaddr*, size_t);
  int windowsEchoServer();
#endif

int main(int argc, char** argv)
{
#if PLATFORM == PLATFORM_WINDOWS

  windowsEchoServer();

#else

  macLinuxEchoServer();

#endif

  puts("Press any key to continue");
  getc(stdin);

  return EXIT_SUCCESS;
}

Mac and Linux Server Setup

/* server.c */
#if PLATFORM == PLATFORM_MAC || PLATFORM == PLATFORM_LINUX

void macLinuxEchoServer()
{
  int socketFileDescriptor;
  struct sockaddr_in serverAddr;
  struct sockaddr_in clientAddr;

  socketFileDescriptor = socket(AF_INET, SOCK_DGRAM, IPPROTO_UDP);

  bzero(&serverAddr, sizeof(serverAddr));

  serverAddr.sin_family = AF_INET;
  serverAddr.sin_addr.s_addr = htonl(INADDR_ANY);
  serverAddr.sin_port = htons(PORTNUM);

  bind(socketFileDescriptor, (struct sockaddr*)& serverAddr, sizeof(serverAddr));

  macLinuxEchoLoop(socketFileDescriptor, (struct sockaddr*)& clientAddr, sizeof(clientAddr));
}

Here we notice the one key difference between the client code and the server code; the use of bind. In a nutshell; from wikipedia “When a socket is created using socket(), it is only given a protocol family, but not assigned an address.”. The assignment of an address is crucial for our sever because without it how else will clients know where to send packets. In other words a socket by itself can identify its protocol from the arguments passed to the socket() creation function but for a process which claims a socket to receive packets based on the networked computers port, there needs to be a one to one immutable relation between IP address and port number.

Mac and Linux Server Echo Loop

Our echo loop is extremely concise, read data from the bound socket, print a notification message and bounce it back to the same client address.

/* server.c */
void macLinuxEchoLoop(int sockFd, struct sockaddr* cliaddr, socklen_t clilen)
{
  int bytesRead;
  socklen_t len;
  char msg[BUFMAX] = {0};

  printf("Waiting for datagrams on 127.0.0.1:%d\n", PORTNUM);

  for(;;)
  {
    len = clilen;
    bzero(&msg, sizeof(msg));

    bytesRead = recvfrom(sockFd, msg, BUFMAX, 0, cliaddr, &len);

    printf("Got message: %s\n", msg);

    sendto(sockFd, msg, bytesRead, 0, cliaddr, len);
  }
}

#endif

Windows Server Setup

On Windows our server setup code is nearly identical to the Mac/Linux version, with the main difference as discussed earlier being the bind() function call.

/* server.c */
#if PLATFORM == PLATFORM_WINDOWS

int windowsEchoServer()
{
  SOCKET sock;
  sockaddr_in serverAddr;
  sockaddr_in clientAddr;
  WSADATA wsaDat;

  int wsaError = WSAStartup( MAKEWORD(2,2), &wsaDat );

  if(!wsaError)
  {
    sock = socket(AF_INET, SOCK_DGRAM, IPPROTO_UDP);

    ZeroMemory(&serverAddr, sizeof(serverAddr));

    serverAddr.sin_family = AF_INET;
    serverAddr.sin_addr.s_addr = htonl(INADDR_ANY);
    serverAddr.sin_port = htons(PORTNUM);

    bind(sock, (struct sockaddr*)& serverAddr, sizeof(serverAddr));

    windowsEchoLoop(sock, (struct sockaddr*)& clientAddr, sizeof(clientAddr));
  }
  else
  {
    return EXIT_FAILURE;
  }
}

Windows Server Echo Loop

At the very minimum the sever echo loop doesn’t need to do much as shown in the code below; add it to server.c and finish it up.

/* server.c */
void windowsEchoLoop(SOCKET sock, sockaddr* cliaddr, size_t clilen)
{
  int bytesRead;
  int len;
  char msg[BUFMAX] = {0};

  printf("Waiting for datagrams on 127.0.0.1:%d\n", PORTNUM);

  for(;;)
  {
    len = clilen;
    ZeroMemory(&msg, sizeof(msg));

    bytesRead = recvfrom(sock, msg, BUFMAX, 0, cliaddr, &len);

    printf("Got message: %s\n", msg);

    sendto(sock, msg, bytesRead, 0, cliaddr, len);
  }
}

#endif

Now go build and run the server.c once that code is running and awaiting new clients go build and run client.c to see the Echo server in action.

Complete Source Code

Click here for Part I