FTP and
HTTP clients are available for Java. Both protocols obviously limit you to the directories that the respective server makes available.
With FTP, you can recursively work your way through all directories, and look for the file you seek.
HTTP is rather limited for searching. You'd have to access pages and hope that you find links in them that point to files, possibly following them recursively. Or maybe the server returns a directory listing when accessing a URL ending in "/". But either way, you'd have to parse the result and interpret it. Not good. A tool like
HttpUnit makes dealing with the resulting HTML easier, giving you a DOM view into it.