The most common way to scale the reads hitting a distributed data store is by adding Read Replicas. These replicas handle all the reads of the systems freeing up the Master to deal with the writes. Although Replicas do help us scale, it brings a new set of problems; and in this essay, we discuss one such issue, called "Read-your-write" consistency, and look at possible solutions.
Nice post to explain an overlooked DB access paradigm. It would be good to mention that “Master Fallback” cannot be used in systems which have upserts/updates.
Nice post to explain an overlooked DB access paradigm. It would be good to mention that “Master Fallback” cannot be used in systems which have upserts/updates.