How does auto partitioning work?

Hi,

I’m wondering how auto partitioning work and how it is superior to the on-prem partitioning?

The documentation says:

This makes client drivers and client programs easier to write and use, and it eliminates partition control as a user-space concern.

Understood, but would auto partitioning also help to support multi-tenancy?

Best regards,

Max

Hi @Maxweber,

Understood, but would auto partitioning also help to support multi-tenancy

In Datomic Cloud partitioning is automatic and it does help support multi-tenancy. The current partitioning automatically distributes across partitons in a DB. The point here is that you do not need to determine a partitioning strategy when using Cloud and Cloud supports multi-tenancy.

Thanks,
Jaret

Hi @jaret,

thanks a lot for your answer. Good to hear that Datomic Cloud was also optimized for multi-tenancy.

May you tell us more about, how Datomic chooses the partition for an entity? We use an attribute :entity/owner to store a ref to the owner entity. Ideally all entities of an owner should end up in the same partition. At the moment we have performance issues with On-Prem Datomic, since we do not use partitioning yet and the peer needs to fetch too many segments from the cold storage, when a customer returns after several hours of inactivity. A migration to On-Prem partitioning would be very difficult for us, since all entity ids would change.

Best regards,

Max