3 ms·
>usecase Dynamo db handles burst traffic (scaling up and down on server count) better than mssql. Dynamo also handles big data better than sql. The last time
by lev99 9y ago
>usecase
Dynamo db handles burst traffic (scaling up and down on server count) better than mssql. Dynamo also handles big data better than sql.
The last time I tried (2015) no amount of indexing or sharding was allowing for decent performance on joins across tables with 300+ million rows in sql server. I've seen dynmodb used in map reduce functions with more data.
Dynamo also doesn't completely suck at typical transcational work, and it's convient to have one datasource across the warehouse and application layers.
- philliphaydon 9y agoWhen I worked in Australia we had a client with ~1.5tb database in SQL Server, largest main table was about 2.5b records. Had no problem returning joined records in low ms times. Wouldn't be good if it had really high traffic but we wrote about ~1-2m records per day into the tables without issue reading it out. SQL Server can perform. But requires a lot of massaging. --- We tried using DynamoDB for logging, not sure why but the team had trouble with document sizes, write limits, in the end it was growing and going to cost more than SQL Server, was replaced with logging to PostgreSQL with much more flexibilty and runs on a t2.micro instance without issue.