My client has a windows application called COPIA that waits for a Microsoft word file to appear in a special input directory and coverts the document to a TIFF file in the output directory.
There is a perl script (called take_this.pl running on another separate windows box) that accepts Microsoft word files from the web page upload and places them (via FTP) in the COPIA input directory and polls the output directory until the TIFF file appears there indicating that COPIA is done with its conversion. Let us suppose we wanted to introduce some redundancy and have a second computer that is running the COPIA application. How can I enhance this take_this.pl perl script so that it (1) checks for the existence of the primary node running COPIA and FTPs the Microsoft word file there (2) checks for the existence of the redundant (and slower) secondary node running COPIA and redundantly copies the Microsoft word file there. (3) Alternately polls each node to see if which is done first (I think I figure this step out myself). What Perl modules might help me here? I know how to enumerate network nodes using a C program but I did not recall such a capability in Stein's "Network Programming with Perl". I suppose there is a perl module to ping a node, I wonder if I would want to use that technique instead of the Microsoft specific approach of enumerating all the nodes on the local network. How can we take this one step further and implement load balancing so that if the primary COPIA computer has a backlog, we copy to the secondary one instead. I guess I can answer my own question here: If the size of the directory of the input for the primary COPIA is double the size of the input directory for the secondary, put it in the secondary one. Is there a better way? Thanks, Siegfried
