[ad_1]
Dispersed units are complicated. Include networking to the blend, and you get traumatic problems like the CAP theorem and Byzantine fault tolerance. Most of these troubles are unidentified to engineers who have to undergo by way of the seller promoting displays, building it hard to establish whether the newest shiny gizmo operates outside of PowerPoint.
I begun amassing articles describing distributed-system gotchas a long time ago, wrote quite a few website posts on the matter in the heydays of the SDN Will Conserve the Environment lemming run, and structured them into the Dispersed Devices Methods website page.
[ad_2]
Source link