How to discover devices on network

  • Thread starter Thread starter stacy
  • Start date Start date
S

stacy

I am developing an application that must look out over TCP on an
ethernet connection to
locate our proprietary devices. The Web application and the devices
are communicating over ethernet with TCP using known IP addresses at
present. We would like to design a method to discover devices in the
system without knowing their IP addresses. Is there a recommended
protocol for this type of problem or do most develop their own
protocol? Any suggestions on how to do this would be appreciated.

Regards, Stacy
 
Hi,

Thanks for the suggestions, we can modify the devices, we are
developing them ourselves. Are you suggesting the I send a message out
to the subnet broadcast address IE: xxx.xxx.xxx.255? I thought that
UDP was not a guaranteed protocol, in which case, how can I be certain
that all devices do indeed receive and reply to the message?

Thanks again,
Stacy
 
Hi,

Thanks for the suggestions, we can modify the devices, we are
developing them ourselves. Are you suggesting the I send a message out
to the subnet broadcast address IE: xxx.xxx.xxx.255? I thought that
UDP was not a guaranteed protocol, in which case, how can I be certain
that all devices do indeed receive and reply to the message?
Click to expand...

'Not guaranteed' does not mean totally unreliable. Very few
broadcasts fail. Still, all you do is broadcast periodically, once a
minute, or once every ten seconds. You will virtually never miss even
one. Missing three or four in a row is essentially impossible!

Of course, some network administrators take a dim view of broadcasts
and don't let them through the switches so make sure that those
broadcasts on that specific port address are acceptable.
 
Back
Top