I love UUID, I hate UUID
At Epsio we’re building a streaming SQL engine which natively integrates with databases/warehouses (both as “source”s and “sink”s). Epsio acts as a stream processing “middle man“- receiving data from one (or more) sources, applying transformations (a...
blog.epsiolabs.com6 min read
Tim De Lange
Been around the block.
Nice writeup. A potential drawback of this is that random access to rows may slow down because of unbalanced b-tree indexes due to the most significant bits being largely the same. It may also have an effect on the distribution of load when sharding big tables across instances.