In my field (computer graphics) most conferences have put clauses allowing authors to host the submitted version of the paper in their websites. Far from ideal, but some compromises from the part of the publishers have been made.