Query Processing in Tertiary Memory Databases
Sunita Sarawagi, Michael R Stonebraker · 1995
Query Processing in Tertiary Memory Databases by Sunita Sarawagi Doctor of Philosophy in Computer Science University of California at Berkeley Professor Michael R. Stonebraker, Chair This thesis presents the design and implementation of a database query processing engine that is optimized for access to tertiary memory devices. Tertiary memory devices provide a cost-effective solution for handling the on-going information explosion. While cheap and convenient, they pose new optimization challenges. Not only are tertiary devices three orders of magnitude slower than disks, but they also have a highly non-uniform access latency. Therefore, it is crucial to carefully reduce and reorder I/O on tertiary memory using effective query scheduling, batching, caching, prefetching and data placement techniques. We make two key modifications to an existing query processing architecture to support such aggressive optimizations: The first is a scheduler that uses system-wide information to make query ...