All Forums Database
sou2011mik 8 posts Joined 01/13
19 Dec 2013
Teradata architechtural behaviour

Hi all,
I would like to know that in which specific scenarios, an AMP will be overloaded
- for example, while firing an untuned query, how would this SQL be submitted to the AMPs and how the AMPs would respond to this untuned query.
 
Kindly let me know the conceptual phenomena taking place.
 
Regards,
Soumik

Raja_KT 1246 posts Joined 07/09
20 Dec 2013

Hi,

Yes, you can think it that way when we have skewness, with respect to data volume, uneven distribution  due to bad choice of PI, bad joins etc.But Teradata system is built for parallelism from scratch and we have to make use of its capabilities.

 

Maybe you can  read about AWT(Amp worker task) to see more into the architecture.

Work00-- Query step coming from dispatcher

Work01-- First level of spawned work

 and the standard reserve pools( 24 reserved AWTs on each amp)

 

Cheers,

Raja

Raja K Thaw
My wiki: http://en.wikipedia.org/wiki/User:Kt_raj1
Street Children suffer not by their fault. We can help them if we want.

sou2011mik 8 posts Joined 01/13
28 Dec 2013

Thanks a tonne Raja!
Could you please provide me a link for such document, so that it can be more understandable to me?
 
Regards,
Soumik

Raja_KT 1246 posts Joined 07/09
29 Dec 2013

Hi Soumik,
For AWT,
you can have a look at this link http://developer.teradata.com/blog/carrie/2010/12/how-to-calculate-your-max-number-of-usable-amp-worker-tasks .  Carrie has given a vivid description  and clarifying many queries.
 
Cheers,
Raja

Raja K Thaw
My wiki: http://en.wikipedia.org/wiki/User:Kt_raj1
Street Children suffer not by their fault. We can help them if we want.

You must sign in to leave a comment.