A cluster is a contiguous allocation of space for one type of database object. Data clusters reduce fragmentation and enable your database to yield better performance from the underlying file system.
Data clusters are specified on a per-area basis. There is one cluster size for all extents in an area. The minimum size of a data cluster is 8 blocks, but you can also specify larger clusters of 64 or 512 blocks. All blocks within a data cluster contain the same type of object. The high-water mark of an extent is increased by the cluster size.
In the Type I architecture, blocks are laid out one at a time. In the Type II architecture, blocks are laid out in a cluster at one time. With the Type II architecture, data is maintained at the cluster level, and blocks only include data associated with one particular object. Additionally, all blocks of an individual cluster are associated with a particular object. In Release 10 of OpenEdge, existing storage areas use the Type I architecture, as do schema. New storage areas can use the Type II architecture or the Type I architecture. To use the Type II architecture, you must allocate clusters of 8, 64, or 512 blocks to an area. If you do not, you will get the Type I architecture. Cluster sizes are defined in your structure definition file. For details on the structure definition file, see OpenEdge Data Management: Database Administration.
