File System and Parse Operations Cache in XML/Directory Import Patches

Hi, Does anyone have any pointers regarding the recurrent import and parse of XML and Directory information? The context is something like a controller that matches a continuously changing string to a list of keys from a XML file or directory import patch, every 4 to 1800 seconds.

If these patches don't implement any cache (for parse or at least read) would it make sense to implement something like a daily import+parse+cache strategy in quartz (with custom bits in javascript)?

My doubts are mainly in memory management issues created by this approach specifically in maintaining state in javascript patches.

Thank you very much for any help! pnbv