An Analysis of "Distributed Free Space" in an Operating and Data Management Systems Environment

Due to subsequent insertion and deletion, the performance of a file will decline as time goes by. In order to maintain a fast response time at the expense of storage space, a technique commonly known as "within distributed free space" has been developed and used in many access methods. In this paper, a procedure to determine and preallocate the size of "within distributed free space" is presented. The optimal block size for a data storage area, which minimizes the CPU operations and I/O interruptions, is also described.