Can a client and a server both use protocol port 5193 on the same
computer at the same time?

Hi jumbie

Regardless of whether you are referring to the TCP or UDP protocols,
there is nothing special about port 5193. However your question seems
to be a little bit confused.

When you start a server process using TCP or UDP, it *listens* on a
particular port. That is to say, it will wait for something to send a
message to that port, at which point it will (hopefully) understand
the message and do something useful, which may well involve sending a
message back to the computer that initiated the connection.

A client connection, however, is simply told the numnber of the port
to connect *to* (and, of course, the machine to contact). It isn't
possible to specify which outbound port to use on the client, you
merely specify which port to connect to at the destination end.

Hence it is certainly possible for a server on a machine to listen on
port 5193, and for a client on the same machine to connect to that
server. In fact this is what the special IP address 127.0.0.1 (usually
aliased as 'localhost') is for. The process is identical to contacting
any other server process running directly on your machine.


Some useful links giving more information:

An introduction to internet protocols:
http://oac3.hsc.uth.tmc.edu/staff/snewton/tcp-tutorial/

A general overview of client-server programming with sockets:
http://www.catalyst.com/support/tutorials/tcpintro/page8.html

Technical, but they're the ultimate authorities:

Internet Standard 0006 / Request For Comments 0768 : UDP
http://www.faqs.org/rfcs/rfc768.html
Internet Standard 0007 / Request For Comments 0793 : TCP
http://www.faqs.org/rfcs/rfc793.html


Regards
iaint-ga


Search queries used:

At http://www.faqs.org/
"transmission control protocol" and "user datagram protocol"

On Google
introduction sockets client server

Yes, this looks good!

Just a minor correction to the above answer:

On many/most platforms, a client most certainly may choose its local
source port.  It is highly recommended that the OS be allowed to give
the client a "random" source port, but that does not mean that the
client cannot try to get a certain one.

For example, see the <a
href="http://www.ibrado.com/sock-faq/#faq37">UNIX Socket FAQ, Section
3 Question 6</a>, which explains the issue in more detail.  Also, see
the <a href="http://java.sun.com/j2se/1.4/docs/api/java/net/Socket.html">Java
API for the Socket class</a> which allows you to bind your client
socket to a certain source port.

As explained in the FAQ referred to above, some application protocols
use the client source port as a sort of authentication mechanism,
though it is a very weak one.  The FAQ mentions rlogind and rshd, and
I believe that ftp might as well when establishing a data connection.