Uber’s blog post about migration from PostgreSQL to MySQL made a lot of buzz in PostgreSQL community. Many of Developers PostgreSQL community realized shortcoming of our table engine (which is the only one yet). As result, many of patches were developer in order to overcome the shortcomings mentioned by Uber. Some of those patches are overlapping, even some of them are in contradictory. Those patches include: indirect indexes (indexes which references primary key value), WARM (write-amplification reduction method), RDS (recently dead storage). Also there are discussions about pluggable table engines and undo log.
In this talk I’ll consider points of Uber’s blog post from PostgreSQL developer point of view. I’ll tell which points I agree, which points I disagree and which points I partially agree. Also I’ll consider developments of PostgreSQL community, and how them can overcome mentioned shortcomings from my point of view.
Alexander is computer scientist (PhD) and developer of Open Source DMBS PostgreSQL (major contributor). In 2015 together with other Russian PostgreSQL developers he founded Postgres Professional company which is Russian vendor of PostgreSQL.
Alexander contributes to PostgreSQL more than 7 years and his contribution includes: pluggable index access methods, advances to GiST and GIN indexes, better algorithms for spatial and fuzzy string search, multicore system optimizations and more.
Company: «PostgreSQL», Russia