a perspective, but you can find more info from the teradata aster website too.
- data is replicated across "WorkerNode" local disk, you can use SAN but...
- data is protected via "rack" aware, Replication Factoring. HDFS like...
- Cluster oriented, commodity hardware may be used and you can even try VM's based approach with Aster Software Only Solution.
- SQL ACID (Postgres) based multiple VWORKER threads based on number of WorkerNodes.
- Full suite of Analytical/Scientifc database implemented functions. C, C++, JAVA , role your own if your a capable.
- Master/Queen Node tracks all WorkerNode processing, health-checks, etc... also peforms key role in aggregations and routing.
Overall my experience has been Positive! Great Discovery Platform....