I hope you are familiar with hadoop, hive, hbase.... programming.You will see MR scripts spawned when you fire queries in hive.With Teradata SQL-H™ you can access Hortonworks Hadoop data say thru table operator, load_from_catalog, using EAH.(I am not sure why the JVM for TD is Oracle 1.6 version. I think 1.5 is fine. Someone can educate me).This is similar to using udfs in query.It pulls data rows and columns from hadoop table, using templeton service, for REST. Data is imported from hadoop nodes into AMPS, parallely.You can join with your TD query. The infos available to PE is then sent to AMP to execute the table operator, importing the data.Hadoop can run in secure (Kerberos..) and in-secure mode. Kerberos set up??? :).
Aster supports both Aster SQL-H and SQL hadoop Hcatalog.With SQL-H we can access unstructured data in hadoop by way of SQL-like interface. There are lots of ready-made functions available to interact with hadoop.Interopability with TD-Aster-- Teradata-Aster Adaptor or utilities. With hadoop--- using Infiniband phsyical connection , connectors or utilities or jdbc. On Aster DB, SQL-MR function can be run to do the search.
With querygrid,heteregeneuous data sources, types, complexities..... are processed and profited by taking to appropriate environments, Aster or Teradata or hadoop to optimize and simplify the processing.
There are lots of benefits behind the scenes. Only when you go to a customer and study their systems, there the benefits can be seen.