In this video Piyush Singh, a Senior Product Manager at Cockroach Labs, explains how you can improve application performance using Duplicate Indexes. Specifically, duplicate indexes create faster reads for your audience. This is an important solution for any business that has an audience in more than one region of the world. The pattern Piyush mentions in the video is scaled to fit the US, but the problems we talk about are even more important in a fully global setting - (the world is just more difficult to draw.)
This video is intended to focus on speeding up reads. But there is also an important survivability component. The pain of old school DBs is even worse if you're replicating data to other geographic regions to improve survivability.
We had a prospect that was using Oracle for this and they would have issues for minutes after users created their accounts where users were not able to log in. You can read that case study here: https://www.cockroachlabs.com/case-studies/top-u-s-financial-software-company-turns-to-cockroachdb-to-improve-its-application-login-experience/ Piyush briefly mentioned fault tolerance while explaining how Duplicate Indexes work in CockroachDB - and it's worth noting that if the 3 nodes in each region are in different "Availability Zones", CockroachDB will be tolerant to failures of individual zones and also entire regions.
If you have further questions about Duplicate Indexes check out this link from our docs: https://www.cockroachlabs.com/docs/v19.2/topology-duplicate-indexes.html#main-content