Let me suggest any other appropriate way (functions or code) to find unique Processor ID.

No, please, let me :)

I assume your desire is to get rid of the dependency on the external binary and do it all in Perl. First step would be CPAN. My first port of call would be Sys::Info, and see whether it builds correctly on your platform.

If it doesn't another approach would be to look at the source to dmidecode, figure out what it does, and write some XS code to duplicate the functionality. This is the approach I used when writing BSD::Process and BSD::Sysctl.

If it boils down to a couple of kernel system calls, you might be able to get away with calling syscall if you can get all the necessary arcana packed up appropriately.

Other than that, there's nothing intrinsically wrong with using an external binary, it's an honourably Lazy approach. You might just want to open it as an input pipe and munge its output directly, rather than fiddling around with grep and sed.

update: *snort* looks like Khen1950fx has the goods. That'll serve me right for not looking at CPAN before posting!