Termin: 25.06.2019, 14:00 Uhr
Ort: Institut für Informationsverarbeitung (13. Stock, Appelstr. 9A, 30167 Hannover)
Abstract: In my talk, i will present MorphStore, a regular in-memory column store database system with a novel compression-aware and highly vectorized query processing concept. Basically, compression using lightweight integer compression algorithms already plays an important role in existing in-memory column stores, but mainly for base data. The continuous handling of compression from the base data to the intermediate results during query processing has already been discussed, but not investigated in detail since the computational effort for compression as well as decompression is often assumed to exceed the benefits of a reduced transfer cost between CPU and main memory. However, this argument increasingly loses its validity even through vectorization which reduces the computational effort. Therefore, we designed and implemented a novel compression-aware and highly vectorized query processing concept that is characterized by the fact that we are able to speed up the query execution by morphing compressed intermediate results from one scheme to another scheme to dynamically adapt to the changing data characteristics during query processing. Our morphing decisions are made using a cost-based approach. Moreover, i will show that MorphStore is well-designed for modern CPUs offering vector extensions as well as for pure vector engines like NEC SX-Aurora TSUBASA.
