Good morning.
We'll be moving to an Exadata architecture in 2 or 3 months. What are the key differences from the developer side of things? E.g. do we see hash joins being comparatively faster? Do we need to worry less about an insert /*+ parallel */ because it's offloaded to the storage nodes?
And are there different details I should look at in ASH or AWR queries? Are there new waits, perhaps equivalent to the RAC concurrency “gc” waits? Are there new stats I should pay attention to?
Are there particular details of RDMA that I will (or should?) notice when I look at system performance?
Thanks all around.