This session will take a deep dive into query scalability with column store indexes and batch mode, this presentation will illustrate how by leveraging vectorised processing and CPU L2/3 cache batch mode scales and compare this to row mode. Stack walking will be used to quantify the cost of conventional page and row compression, column store versus the new to SQL 2014 column store archive compression and row mode versus batch mode operators. The affect of storage that can and cannot keep up with the available CPU resource will be covered along with how well batch mode scales across 24 schedulers.
Presented by Chris Adkin at SQLBits XII
Slide Deck 1.4 MB
MP4 Video HD 499 MB