Home Page for the TeradataForum
 

Archives of the TeradataForum

Message Posted: Wed, 03 Mar 2010 @ 08:16:03 GMT


     
  <Prev Next>  
<<First
<Prev
Next> Last>>  


Subj:   Data Skew vs Single Amp
 
From:   Barrineau, Claybourne

We are working on a tactical application and want to ensure that all queries running against our single Fact Table will query via a single AMP.

Unfortunately, the only way to guarantee this in all ~300 cases is to create a PI which unevenly distributes the data (some AMPs having 10x as many rows as other AMPs.) In the most extreme cases, we have up to 50,000 rows per hash value. On the other end of the spectrum, we have as few as 12 rows per hash value.

Does anyone have a recommendation regarding the point at which the downside of Data Skewness will trump the benefits of Single Amp Access for a tactical workload? Our system is a 15 node 5555H. We need to prove that we can run 3000 'concurrent (with a grain of salt)' requests thus the need to guarantee single amp access.


Thanks in advance,

Clay

Claybourne Barrineau
The Home Depot - Teradata Architect



     
  <Prev Next>  
<<First
<Prev
Next> Last>>  
 
 
 
 
 
 
 
 
  
  Top Home Privacy Feedback  
 
 
Copyright for the TeradataForum (TDATA-L), Manta BlueSky    
Copyright 2016 - All Rights Reserved    
Last Modified: 27 Dec 2016