Understanding Multi-Region Application Architecture

Conclusion

In this report, we examined techniques to consider and pitfalls to avoid when building applications that span multiple geographical regions. In the previous chapters, we created a distributed application using just a handful of modern tools that will provide users across multiple continents a great experience while containing absolutely zero percent cruft, spaghetti code, or config creep.

In the process, we saw that multi-region deployments don’t have to be expensive. In fact, they can often be more cost-effective than deploying isolated applications across those same regions. Fewer isolated components means less operational overhead and lower cost. With a distributed SQL database (and horizontally scalable software in general), you gain not only failure resilience but also the ability to perform maintenance without downtime.

Multi-region application architecture lets us operate a single logical database across many distributed geographical areas. Recall how that US mobile gaming company was able to use CockroachDB as a single source of truth for geo-fencing user data according to where a bet was placed, satisfying data locality regulations. Rather than distributing themselves with a database in each geographic location (and hiring 5 to 10 engineers to run each one), they had a distributed system with centralized management. This is a compelling example of how multi-region application architecture, paired with the right tools, can save businesses from redundant infrastructure ...