Host: The Japan Society for Management Information
We have discussed disk allocation methods for Cartesian product files by introducing error correcting codes, and have clarified the performance of the methods by system evaluation models using rate distortion theory. Let us assume qn Cartesian product files with n attributes and store qn files into g (<qn) disks. For a partial match query(access request), we propose new disk allocation methods which are able to access in parallel to the disks as much as possible. Then we show that the proposed methods exhibit good property when n becomes large.