I would also be up for some reading regarding optimizations/architecture choices for a 10:1::write:read setup.