Why Is Data Consistency a Challenge with Horizontal Scaling?

When you scale vertically, you're just upgrading the hardware to a more powerful single machine that hosts all your data, so everything stays in sync naturally. But with horizontal scaling, you distribute your data across multiple machines or locations. The internet isn't 100% reliable, and if two instances are far apart, it can be tough to keep everything updated at the same time. If someone makes a change on one instance, there’s always a delay before it reflects on another instance. That inconsistency can lead to issues where you might get outdated data when querying, which is something you don't really worry about when everything's on one server.

Exactly! And to add to that, the syncing issues multiply the more instances you have. Each instance might have its own copy of the data, and ensuring that each one stays updated becomes a complex task. If the connections between these instances aren't great, there can be significant delays. Even with sharding, if the data isn't synced properly, you might end up querying one database and getting outdated information that doesn't match what's in another database.