Netezza performance is something that is out of this world, for someone like me coming from a T-SQL and relational background. I think even if you had considerable BI experience with CUBES and ETL processes your experience would not be considered complete if you haven't dealt with data in the terabytes.
I think i am a bit lucky to be working on this.
Consider some of the technology behind Netezza server. It is a server rack with 28 computers each having its own memory, sql application, processor. Each server has a fail over mechanism for an alternate server.
It has a concept of bringing the query to the data, that means that query is not loaded into memory but fetched directly from the disk, so i guess that would mean A big okay for table scans and no Indexes?
Rather than moving data into memory or across the network for processing, filters records as they stream off the disk, delivering only the relevant information for each query.
Netezza also uses F.P.G.A (Field programmable Gate Array) at the processor level. So this means you can flash the processor with a snippet of code and it then gets optimized since it gets built as a machine language instruction.
This to me is the coolest concept i have ever heard of when it comes to DW servers.
