Data Morphing: An Adaptive, Cache-Conscious Storage Technique Richard A. Hankins and Jignesh M. Patel   View Paper (PDF)   Return to Caching (Session B1) Abstract The number of processor cache misses has a crit- ical impact on the performance of DBMSs run- ning on servers with large main-memory configu- rations. In turn, the cache utilization of database systems is highly dependent on the physical or- ganization of the records in main-memory. A re- cently proposed storage model, called PAX, was shown to greatly improve the performance of se- quential file-scan operations when compared to the commonly implemented N-ary storage model. However, the PAX storage model can also demon- strate poor cache utilization for other common op- erations, such as index scans. Under a workload of heterogenous database operations, neither the PAX storage model nor the N-ary storage model is optimal. In this paper, we propose a flexible data stor- age technique called Data Morphing. Using Data Morphing, a cache-efficient attribute layout, called a partition, is first determined through an analysis of the query workload. This partition is then used as a template for storing data in a cache- efficient way. We present two algorithms for com- puting partitions, and also present a versatile stor- age model that accommodates the dynamic reor- ganization of the attributes in a file. Finally, we experimentally demonstrate that the Data Morph- ing technique provides a significant performance improvement over both the traditional N-ary stor- age model and the PAX model. ©2004 Association for Computing Machinery