My previous post discussed the sharding conflicts often associated with a distributed relational database. Sharding conflicts are by no means the sole challenge of relational database distribution, however, they do demonstrate how distributed databases can become quite complex if not handled correctly. Today, I would like to look at how we can reduce and eliminate sharding conflicts: by storing together data that will be used together. However, some tables that relate to the chosen data distribution (shard) key do not possess the key. At ScaleBase, we call these tables “cascaded tables.”
Let me explain what they are, and how we can use them to eliminate sharding conflicts.