[cordite]

AFAIK, Datomic doesn't have the same kind of distributed load management out of the box for you, where you can do analytical queries over the whole data set.

Datomic is designed such that load from analytic queries is local to the quering machine (apart from storage retrieval which can be cached and replicated), thus it does not need to be ran on a shadow server or some distanced system from the live production system.

Because the data is immutable in Datomic, there is no need for worrying about locks on tables or documents for contention of future writes--since it reads the data at the time of the query starting from immutable files (joined with the database transaction 'novelty' buffer/log since the last database indexing operation).

It is also up to the querying machine to store the conclusions from such queries in whatever way they like. (If this means going to another datomic instance, or put on HDFS or GFS, by all means.)