Dispersed methods are difficult. Add networking to the mix, and you get traumatic difficulties like the CAP theorem and Byzantine fault tolerance. Most of those problems are unknown to engineers who have to endure as a result of the vendor advertising and marketing displays, generating it hard to decide whether or not the hottest shiny gizmo works outside of PowerPoint.

I started off collecting posts describing dispersed-technique gotchas decades back, wrote quite a few blog posts on the topic in the heydays of the SDN Will Help you save the Entire world lemming operate, and structured them into the Dispersed Methods Sources page.