Comments on: Azure SQL Data Warehouse Costs vs AWS Redshift

By: Poem

Poem — Fri, 11 Oct 2019 06:44:14 +0000

Doing comparisons just on configuration is hard and as always there are technical differences behind the scenes. If you try to compare apples with apples on the config side you end up with nearly the exact same costs for SQL DW and Redshift (compare on demand price for DW400c vs. dc2.8xlarge -> 5.56 vs 5.6 USD per hour). You should compare Price/Performance using an independent benchmark like TPC-H where it seems that Azure SQL DWH has an advantage over Redshift: https://gigaom.com/report/data-warehouse-in-the-cloud-benchmark/

By: Raja K Thaw

Raja K Thaw — Thu, 05 Sep 2019 03:32:47 +0000

For Azure, DWU, cDWU definition and values itself are blurred. How do we factor with actual size of data. DC2,DS2 are clear for AWS and China region has NVMe SSD. PostgreSQL columnar features are clear and powerful. Eaxmple, each DS2.XLARGE compute node has two slices, whereas each DS2.8XLARGE compute node has 16 slices, so that we can leverage MPP architecture. I have not seen much how Azure SQL DWH leveraging the columnar features , though it says columnstore. But AWS Redshift is clear columnar and it is buttressed by facts on vacuum and analyze. Of course the list of comparison goes on , if we see the technical part of it.

By: Thomas LaRock

Thomas LaRock — Sat, 23 Feb 2019 17:53:33 +0000

In reply to Nnanna. Thanks for the comment. You can also delete the Redshift cluster, stopping all charges. While not the same as the pause available in Azure SQL DW, it serves the same purpose. You would redeploy Redshift from a snapshot and be back up and running in minutes. I chose the 730 hours as a way of keeping the comparison fair in terms of understanding the costs for each service.

By: Nnanna

Nnanna — Fri, 22 Feb 2019 10:19:05 +0000

The compute hour for Azure SQL Data Warehouse is 730 hours which is continuous computation for a month. However, compute and storage are separate which means you can pause compute when not need which further reduces the cost ofr Azure SQL DW.

By: Thomas LaRock

Thomas LaRock — Tue, 04 Sep 2018 19:26:24 +0000

Your example of $3864 is for 1 node, but Azure has 2 nodes in their cost. My post even mentions this. I like your 13 node example, but to make a fair comparison we would need to find similar smaller machines for Azure, and they don’t offer similar hardware.

By: Yangliang Li

Yangliang Li — Wed, 29 Aug 2018 09:16:56 +0000

If you want to store 2TB data in AWS Redshift, you can choose 1 node of dc2.8xlarge, which gives 2.56TB SSD and costs $3864.96 per month; or choose 13 nodes of dc2.large which gives 2.08 TB SSD and costs $2616.9 per month. Both are much lower than Azure SQL Data Warehouse’s $5789.14.

By: Thomas LaRock

Thomas LaRock — Tue, 05 Jun 2018 03:21:53 +0000

Is that pricing for 2 nodes, or just one? The example above shows 2 nodes. I tried to compare the 2 nodes in AWS to the SQL DW compute option. And, as my posts have suggested, trying to do comparisons between the services is not easy. I’ve done my best to be objective and fair when comparing services and prices.