RE: Oracle Exadata Machine
Date: Wed, 13 May 2009 12:31:48 -0500 (CDT)
Don't know about Exadata but I don't think Netezza uses indexes at all. Or at least it didn't at one time.
This was explained to me by way of a joke, which I'll keep short.
A guy shows up to play golf and his partner is there with a Gorilla. He says,
"I'll bet this Gorilla can beat you"
The Gorilla tee's up and hits a drive 500 yards straight down the fairway. The guys admits defeat and pays up. He then asks, "So, how does he putt" and the guy with the Gorilla says "Same way, 500 yards".
So, that's how it works with Neteeza. Every query is a full table scan and the time is essentially the same regardless of complexity. The idea is that the full table scan can be done very fast.
At least that's what they claim. I don't want to sound like a salesman and don't have any real world number. My only point is that after looking at the two architectures, I like theirs better. But, if you need a certain level of Oracle compatibility then that's a different story.
> I've asked Kevin a few times and haven't got an answer but I've wondered how
> exadata does with frequent single keyed reads.