I have no big experience with spiders, but don't you think that the web will be the bootleneck?

I expect Perl to parse HTML much faster than the server responses come in.