Firstly, I have been using 64-bit Perl more or less exclusively for the past few years on a variety of Linux, Windows, and Mac systems, with 50+ modules (many of them XS, don't have an exact count on that, though), and have had zero issues that would have been avoided with a 32-bit Perl.

Secondly, my own anecdotal experience shouldn't mean squat to you. :-) Fortunately, the Perl module ecosystem (and Perl itself) are blessed with some of the best unit tests in the industry (see Test::More for a start). As long as your own software also has good unit tests, you are already a long way to verifying your installation.

Thirdly, if you still have any doubt, it's worth questioning whether 64-bit Perl would actually make any substantial difference in your installation. By themselves, 64-bit programs do not inherently outperform their 32-bit counterparts. There are a couple of cases where they do:

• (The big one): Your program needs to access more than ~3GiB of RAM, go with 64-bit
• 64-bit architectures have additional registers that can (at the C compiler level) be used to optimize some compute-intensive routines. Benchmarking is the only sane way to tell whether this will be worthwhile.

However, there's a downside: 64-bit programs are larger (again, depends, but I've seen 10-30%), and consequently use more RAM, which can increase memory traffic, which can actually degrade performance a little. Probably not terribly significant, but again, benchmark.

For "SAN Management Servers", you are probably not using significant RAM or doing anything computationally expensive (think: decoding human genome, solving chess, etc.), so to be honest, it's probably not going to make a huge difference. So, as above, unit tests (and your own functional and integration testing) will guide you.