If you're only spiking to %50 adding more cores isn't a great suggestion as a starting point. Unfortunately troubleshooting performance issues are not a simple answer, there are many possibilities....

NULL is unknown, therefore you cannot evaluate an equation to it. Let's say X is 2.

X=2 is True, X=3 is False, X=NULL is unknown, it isn't True or False...

The NC indexes do have the clustering key, but it is still a lookup on the clustered index to retrieve the data from the leaf pages. It is similar to...

I agree with Pablo, but it also depends on how you are handling slowly changing dimensions, if at all.

I personally only use the natural key(s) in the staging and SK's...

WITH A as

(SELECT (name+cast(id as char)+xtype) test, row_number() over(order by name asc) rn

FROM master..sysobjects)

SELECT * FROM a where rn = 3 --choose whatever row number you want

-Jake

Your first option looks great.

If you have the ability, I would also recommend:

1) Use RAID 10 on the log file drive

2) Use RAID 5 or 6 on the data files...

I agree with Jeff, ETL is in no way specific to the dimensional data warehouse. I define ETL as any process which takes data from one source into another source....

Grasshopper,

Just to throw it out there, you do have the option "Allow remote connections to this server" enabled right?

I have seen several solutions to this error, some easy, some incredibly...