1

I have a web application which currently relies on MySQL. I have it running on AWS at the moment (using a stock RDS) and it can tolerate about 500 requests a second (with a 50/50 mix of INSERTS and SELECTS) and under 1 second response time. However, I am concerned if this number increases to 1000 for example, the RDS will simply not be able to keep up and response times will increase. To reduce this, I've created a "Read Replica" for performing reads to reduce the load on the "Master" server.

I was wondering however, if it was worth me setting up my own MySQL Cluster instead. Would this give me true scaling "out"? Could I simply fire up more nodes when my MySQL cluster was getting busy etc etc? Would it remove the need for "Read Replicas"? Does anyone have any experience with this on AWS?

Thanks in advance for any advice!

Schodemeiss
  • 230
  • 1
  • 2
  • 7

1 Answers1

0

What is the instance type that you are using for your RDS? Have you read the FAQ here -- http://aws.amazon.com/rds/faqs/#19

If you are already using the largest instance type see here --

Q: How can I scale my DB Instance beyond the largest DB Instance class and maximum storage capacity?

Amazon RDS supports a variety of DB Instance classes and storage allocations to meet different application needs. If your application requires more compute resources than the largest DB Instance class or more storage than the maximum allocation, you can implement partitioning, thereby spreading your data across multiple DB Instances.

Hope this helps

Chida
  • 2,491
  • 1
  • 17
  • 29
  • I assume what they refer to as "partitioning", they actually mean Sharding? – Schodemeiss Aug 20 '12 at 13:11
  • Thy are slightly different. Partitioning is physical split of the same logical database. Sharding is logical split of the database (can be physically split as well) – Chida Aug 20 '12 at 13:21