CST363: Week 5 (Week 21)

 Mark Winand, the author of the site, "Use the Index Luke," has a page entitled, "slow indexes."  

Winand explains that "slow indexes" is an incorrect descriptor that is used due to misunderstanding how index operations work. In general, we know that for large databases that don't have write-heavy workloads, well designed indexes tend to speed up queries. Winand points out that many of those who attribute slow indexes to unbalanced or malformed trees aren't taking into account that index lookup is constitued by three steps: 1) tree traversal, 2) following the leaf node chain, and 3) extracting the data from the table. Rather, they incorrectly believe that the index look up only includes step 1 (tree traversal).

Winand is just explaining that "slow indexes" is a misnomer. Indexed database queries are faster when deployed for the right applications. The shortcoming is that we aren't always clear on what the underlying process or technology is doing and make incorrect assessments. It's a good example of a case where it pays to look under the hood and take a closer look.