How to measure performance?

kduke · Post by **kduke** » Wed Apr 05, 2006 1:08 pm

We decided it might be a legal hassle. If you posted IBM always beating Sun even though the servers were not the issue. Maybe they had EMC drives and the disk drives caused these performance gains.

How do you compare apples to apples? EtlStats along with hardware profiles could produce good stats. We are at the mercy of the users being honest.

Not sure how many EtlStats users are willing to share this information but I would be open to hosting it. Let me know if you are interested by posing to this thread. Maybe we should do a poll.

DSguru2B · Post by **DSguru2B** » Wed Apr 05, 2006 2:11 pm

I wouldn't mind doing so!
Definetly there needs to be a poll set up.
And my doubt more than half of the users don't even know what you are talking about. I would suggest to rephrase the idea. Just a suggestion.

gateleys · Post by **gateleys** » Wed Apr 05, 2006 3:16 pm

Kim,
Count me IN. Would be really nice to know where I stand, and get or share possible suggestions for better performance. My numbers may raise eyebrows in certain instances, but then what the heck!! That's why I want to share it with you guys...and improve.

gateleys

ray.wurlod · Post by **ray.wurlod** » Wed Apr 05, 2006 4:23 pm

jusami25 wrote:I'm looking for any numbers or a kind of benchmark comparison for my rows/second number.

As I have posted many times, rows/sec is an almost meaningless metric. It's only valuable if you can guarantee homogeneous row sizes and you can guarantee that "all else is equal". It never is.

Prefer elapsed time, since the KPI is usually the ability to meet a particular load window.

If you MUST have a rate, use something like MB/minute.

kduke · Post by **kduke** » Wed Apr 05, 2006 5:13 pm

EtlStats has MB/Sec now.

DSXchange

How to measure performance?

Re: Has been The performance statistic report built??