Currently, when a table has indexes, the row is first inserted into base table and then the corresponding rows are added into the indexes sequentially. Hbase operations are blocking. Hbase operations are done in a different thread making it non-blocking. This allowed the Trafodion engine to flow the tuples through other operators enabling other operators to be scheduled and the tuples processed by them. This allowed the Trafodion engine to initiate insert into base table and indexes in parallel.