Hyper-replicated collections and sharded collections
There are two distribution mechanisms for collections:
Hyper-replicated collection - a collection is present on every computational node where database is scheduled. This distribution mechanism is suited for small collections, which will be effectively joined with sharded collections by using JOINs.
To create hyper-replicated collection, select "Enable hyper replication for this collection" option under "Advanced" section.
Sharded & replicated collection - a collection is split among many computational nodes to divide storage requirements, parallelise processing and replicated for high availability. This distribution mechanism is suited for large collections to ensure fast query processing.