The Modified Object Buffer: A Storage Management Technique for Object-Oriented Databases

Object-oriented databases store many small objects on disks. Disks perform poorly when reading and writing individual small objects. This thesis presents a new storage management architecture that substantially improves disk performance of a distributed object-oriented database system. The storage architecture is built around a large modified object buffer (MOB) that is stored in primary memory. The MOB provides volatile storage for modified objects. Modified objects are placed in the MOB instead of being immediately written out to disk. Modifications are written to disk lazily as the MOB fills up and space is required for new modifications. The MOB improves performance because even if an object is modified many times in a short period of time, the object has to be written out to disk only once. Furthermore, by the time an object modification has to be flushed from the MOB, many modifications to other objects on the same page may have accumulated. All of these modifications can be written to disk with a single disk write. This thesis evaluates the modified object buffer in combination with a number of disk layout policies that make different tradeoffs between read performance and write performance. The MOB has been implemented as part of the Thor object-oriented database. Results from simulations and from this implementation show that the MOB improves the performance of a read-optimized disk layout that preserves clustering by over 200%. As a result, the read-optimized disk layout consistently out-performs a write-optimized disk layout that does not preserve clustering. The readoptimized disk layout provides better overall performance on a wide range of workloads, including workloads that write a lot of data. Performance results also show that for typical object-oriented database access patterns, the modified object buffer architecture out-performs the traditional page based organization of server memory used in many databases and file systems.