Using Sockets to send and receive data

Question

I am using sockets to connect my Android application (client) and a Java backend Server. From the client I would like to send two variables of data each time I communicate w

Answer 1

I assume you are using TCP sockets for the client-server interaction? One way to send different types of data to the server and have it be able to differentiate between the two is to dedicate the first byte (or more if you have more than 256 types of messages) as some kind of identifier. If the first byte is one, then it is message A, if its 2, then its message B. One easy way to send this over the socket is to use DataOutputStream/DataInputStream:

Client:

Socket socket = ...; // Create and connect the socket
DataOutputStream dOut = new DataOutputStream(socket.getOutputStream());

// Send first message
dOut.writeByte(1);
dOut.writeUTF("This is the first type of message.");
dOut.flush(); // Send off the data

// Send the second message
dOut.writeByte(2);
dOut.writeUTF("This is the second type of message.");
dOut.flush(); // Send off the data

// Send the third message
dOut.writeByte(3);
dOut.writeUTF("This is the third type of message (Part 1).");
dOut.writeUTF("This is the third type of message (Part 2).");
dOut.flush(); // Send off the data

// Send the exit message
dOut.writeByte(-1);
dOut.flush();

dOut.close();

Server:

Socket socket = ... // Set up receive socket
DataInputStream dIn = new DataInputStream(socket.getInputStream());

boolean done = false;
while(!done) {
  byte messageType = dIn.readByte();

  switch(messageType)
  {
  case 1: // Type A
    System.out.println("Message A: " + dIn.readUTF());
    break;
  case 2: // Type B
    System.out.println("Message B: " + dIn.readUTF());
    break;
  case 3: // Type C
    System.out.println("Message C [1]: " + dIn.readUTF());
    System.out.println("Message C [2]: " + dIn.readUTF());
    break;
  default:
    done = true;
  }
}

dIn.close();

Obviously, you can send all kinds of data, not just bytes and strings (UTF).

Note that writeUTF writes a modified UTF-8 format, preceded by a length indicator of an unsigned two byte encoded integer giving you 2^16 - 1 = 65535 bytes to send. This makes it possible for readUTF to find the end of the encoded string. If you decide on your own record structure then you should make sure that the end and type of the record is either known or detectable.