Optimal interval partitioning for temporal databases

We analyse the problem of partitioning a collection of intervals into two or more subcollections, each of them holding intervals that intersect with a certain range of the intervals' domain. Intervals that intersect with more than one range impose a considerable overhead on processing and indexing interval data. The goal of interval partitioning (IP) is to find partitions that minimise the number of overlapping intervals. We describe an algorithm for finding optimal partitions and prove its correctness. An example scenario that is used throughout the paper gives an idea of the benefits of optimal interval partitioning.

[1]  Thomas Zurek.  Parallel Temporal Joins , 1997, BTW.

[2]  Arie Segev,et al.  Efficient Indexing Methods for Temporal Relations , 1993, IEEE Trans. Knowl. Data Eng..

[3]  Richard T. Snodgrass,et al.  The TSQL2 Temporal Query Language , 1995 .

[4]  H. B. Zhou,et al.  Two-Stage m-Way Graph Partitioning , 1993, Parallel Comput..

[5]  Sushil Jajodia,et al.  Temporal Databases: Theory, Design, and Implementation , 1993 .

[6]  Ramez Elmasri,et al.  The Time Index and the Monotonic B+-tree , 1993, Temporal Databases.

[7]  llsoo Ahn,et al.  Temporal Databases , 1986, Computer.

[8]  Nikos A. Lorentzos,et al.  SQL Extension for Interval Data , 1997, IEEE Trans. Knowl. Data Eng..

[9]  Margaret H. Dunham,et al.  Join processing in relational databases , 1992, CSUR.

[10]  Curtis P. Kolovson.  Indexing techniques for historical databases , 1989, [1989] Proceedings. Fifth International Conference on Data Engineering.

[11]  Thomas Zurek,et al.  Optimisation of Partitioned Termporal Joins , 1997, BNCOD.

[12]  Christian S. Jensen,et al.  Efficient evaluation of the valid-time natural join , 1994, Proceedings of 1994 IEEE 10th International Conference on Data Engineering.

[13]  Richard R. Muntz,et al.  Temporal Query Processing and Optimization in Multiprocessor Database Machines , 1992, VLDB.