I would split this table and create hash cluster for each table, but you have to find the spliting rules, one of join conditions in sql query has to based on hash value, you have to find a good hash function. Then create view which join all parts of the table and trigger on the view which makes DML operation on this tables. Don't forget that you may make indexes on this table for increasing responce time. That is my opinion.

Best wishes!
Dmitri

Originally posted by CeciMego
sreddy, kgb thanks for the replies!
the table is around 26 Gb.. and it's actually mandatory to split it, so my main concern is wether to have them all in one cluster, or create a cluster for each table