Consistent Hashing vs Direct Mapping

[espresso maker]

Hi,

Sorry this is not a a direct SQLAlchemy question but I thought of giving it a shot.

I am trying to decide whether I go with consistent hashing or direct mapping to distribute server logs over a set of databases.

I can add a column to the servers table that points to the database that will store the server logs

vs

I can use a Consistent Hashing Ring that can map server_id to a database that will store the server logs.

The advantage of manually assigning a server to a log db instance is that I can distribute the logs to new servers efficiently (least loaded) and can avoid having a particular log db instance getting assigned new servers to it if it reached its capacity.

Any thoughts of either approach?