Instead of using the shared flag would it be better to use Thread::Queue::Monitored as I have found that the array loading is pretty much a drag and am thinking of spawning the threads, loading the ips directly to the queues and at the end of the queing fire off a series of exit commands equal to the threads waiting.

I am trying it now and hopefully have some code soon as an example, but was just curious if you had tried it?

Update:
It appears that Thread::Queue::Monitored only sets up one thread to Monitor the queue which I am going to try to use as the back end process to get the results of the pings instead of trying to shove them into the array. Have the Monitored Queue do the db processing.

No, I have not tried using Thread::Queue::Monitored yet... But I will look into it, thanks.

This is the code I have come up with. This is for a Win32 machine and I am using Win32::ODBC with MySQL. I have not coded it yet to be database independent but that will happen as I update the code but the thread processing works fine and I am able to scan a class A fairly efficiently. There seems to be a small memory leak in it but it does not affect the scan adversely. Memory grows very slowly over a large period of time. With the command line options though you can have the script scan ranges from a batch file or bash script exit and than do another range minimising the affects of the leak.

The Pod is not complete as I am still working on the script. I have to remove references to cab extract as I am doing that process in another script and I have to describe all command line options though it is fairly evident from the code what the options do.

One major note is we were getting responses where the ping replied by gethost returned null. I created a unique computer name based off that and am currently working on a clean up section that tries to identify the invidual devices.

Any additional clean up or efficiency suggestions would be appreciated.