RDS ≠ Aurora: The costly mistakes companies make when switching from Amazon RDS to Aurora

Описание: When migrating from Amazon RDS to Aurora, many companies make the costly mistake of assuming they behave the same way in terms of I/O and storage allocation.

However, the way RDS and Aurora handle storage and I/O is totally different, which can lead to a much higher cloud bill if not properly understood.

With RDS, you pay for the allocated storage capacity, regardless of how much you actually use. The I/O performance is linked to the storage size you allocate, so to get more I/O you need to provision storage you may not need for your data.

In contrast, with Amazon Aurora:

➡ You only pay for the actual storage you use, not allocated capacity. Aurora's storage automatically grows and shrinks in reaction your needs.

➡ I/O is dependant on the workload you run and is not limited in the same way as RDS allowing you to run at full speed without I/O throttling.

➡ By default, I/O is billed per 1 million operations on Aurora whereas on RDS it's included in the storage price.

So, if you migrate from RDS to Aurora and you may end up with significantly higher I/O, leading to much bigger bills.

In addition, RDS I/O does not directly translate to Aurora I/O, so even if you take your RDS I/O metrics and plug them into the AWS Cost Calculator for Aurora you may still find your actual bills are significantly higher!

This is why I/O optimization is so important with Aurora.

This clip comes from episode 15 of Cloud Masters, where we had two AWS database specialists cover Aurora I/O optimization, moving from RDS, and design principles for Aurora databases.

Watch the full episode here:    • Amazon Aurora Deep Dive: Design, migration...  

Need personalized help with your Amazon RDS/Aurora, or with your cloud infrastructure more generally? Let's talk about how DoiT can help your team optimize performance, reduce costs, and accelerate your cloud journey. Reach out here: https://www.doit.com/contact/

#database #aws #cloudcomputing