in reply to What if mod_perl is not an option?

I suggest that you do an experiment to see how your site is going to scale. Set up test cases to exercise a database loaded with fake data, and create a dummy mock-up of your content.

Use LWP to emulate various levels of user load on your site. Create a graph that shows page-load time versus the number of users. This is a great way to evaluate different coding approaches and modules, not just mod_perl.

You could also create your own little network to experiment with the performance improvement that you expect to get from mod_perl.

It's great to be able to go into a meeting and say something along the lines of, "We can serve 500 users with the CGI setup, but when we start to approach this limit, let's reconsider and evaluate a mod_perl approach, which I think will serve at least 5000 users. After all, I was able to emulate 3000 users on a P300 that I picked up at a garage sale."

It's also fun to offer load testing services for competing approaches! I have offered this several times, and no one has ever taken me up on it.

It should work perfectly the first time! - toma