SQL Azure Performance Benchmarking

Question

Post reply

SQL Azure Performance Benchmarking

Paul Brewer

SSCrazy

Points: 2955
More actions
February 4, 2016 at 12:08 am

#399671

Comments posted to this topic are about the item SQL Azure Performance Benchmarking

Viewing 9 posts - 1 through 8 (of 8 total)

You must be logged in to reply to this topic. Login to reply

Robert Sterbal SSChampion Points: 11050 More actions · Answer 1

Very nice technical presentation.

Did you do a cost comparison or model as well?

412-977-3526 call/text

Paul Brewer SSCrazy Points: 2955 More actions · Answer 2

That's a really good idea, I should have considered cost too instead of focusing completely on the performance perspective. Factoring in cost would added a dimension the articles lacking, trying to identify the optimum performance / price .

Thanks for the comment/suggestion

Regards

Paul

brent 28809 SSC Rookie Points: 35 More actions · Answer 3

Nice article and very useful.

Does this test consider the "disk" IO and sustained reads/writes (as you would have in a DW/BI platform)?

Recently we have moved all of our SQL / SQL BI into Azure - with our support vendor deciding A series was sufficient for BI, which it was if you did not consider disk IO - to cut a long story short we have ended up with DS13 over P30s just to get back to a usable DW - this configuration simply sucks your operational budget.

Paul Brewer SSCrazy Points: 2955 More actions · Answer 4

The disk io associated with this script is almost exclusively writes. It clears down data at the start of each call, inserts a starting dataset then generates new patterns. This makes performance consistent but it isn't a good simulation of a DW load. To be honest I'm not an Azure expert either and aren't very familiar with all the different scaling options. This article was my 1st brush with Azure, I can't offer much advice though than it seems great.:-)