Multinode architectures allow users to write data to multiple nodes concurrently. There are two main categories: tightly coupled and loosely coupled:
- Tightly coupled database clusters: These allow a single image of the database, so there is less perception that you're even connected to a cluster at all. This consistency comes at a price—the nodes of the cluster cannot be geographically separated, which means if you need to protect against site disasters, then you'll need additional technology to allow disaster recovery. Clustering requires replication as well.
- Loosely coupled database clusters: These have greater independence for each node, allowing us to spread out nodes across wide areas, such as across multiple continents. You can connect to each node individually. There are two benefits of this. The first is that all data access can be performed quickly against local copies of the data. The second benefit is that we don't need to work out how to route read-only transactions to (a) standby node (s) and read/write transactions to the master node.