2.3 Automate Schema Design for Large Scientific
Databases Using Data Partitioning
As discussed by Stratos Papadomanolakis and Anastassia
Ailamaki, who investigated to Automate Schema Design
for Large Scientific Databases Using Data Partitioning,
"To optimize performance, database researchers have
proposed data placement and partitioning schemes [4][5].
Vertical partitioning is known to optimize I/O performance
since the early days of relational databases [6].
They propose AutoPart, an algorithm that automatically
partitions database tables utilizing prior knowledge of a
representative workload. AutoPart suggests an alternative,
high-performance schema that executes queries faster than
the original one and can be indexed using a fraction of the
space required for indexing the originalschema.