Optimizing OID Indexing Cost in Temporal Object-Oriented Database Systems

In object oriented database systems OODB with logical OIDs an OID index OIDX is needed to map from OID to the physical location of the object In a transaction time temporal OODB the OIDX should also index the object versions In this case the index entries which we call object descriptors OD also include the commit timestamp of the transaction that created the object version In this report we develop an analytical model for OIDX access costs in temporal OODBs The model includes the index page bu er as well as an OD cache We use this model to study access cost and optimal use of memory for index page bu er and OD cache with di erent access patterns The results show that the OIDX access cost can be high and can easy become a bottleneck in large tempo ral OODBs the optimal OD cache size can be relatively large and the gain from using an optimal size is considerable and because access pattern in a database system can be very dynamic the system should be able to detect this and tune its memory use according to this The cost models in this report can be of valuable use for optimizers and automatic tuning tools in temporal OODBs The primary context of this report is OID indexing in a temporal OODB but the results are also relevant in the context of general secondary index access cost and index entry caching