The post Death match! EBS versus SSD price, performance, and QoS on the MySql performance blog has an impressively detailed performance analysis on storing data on SSD’s versus Amazon EBS and concludes:
In conclusion, in a knock-down-drag-out fight between the EBS gang and the SSD thugs, a small number of SSDs mops the floor with the competition, and walks away with the prize.
They failed of course to notice that the EBS gang didn’t bother pitching up and rather went looking for girls at the local pub. Regardless of Amazon’s role in the metrics, concluding that direct attached SSD’s are faster than networked attached storage should come as no surprise to anyone, regardless of how you try and drape it in price/performance figures that the author did.
What is interesting is the number of commenters that found this information useful and more importantly illustrates a complete lack of understanding by the MySql community on how Amazon EC2 and its related offerings work. Despite the fact that on AWS MySql databases should be on RDS, the point of cloud computing architectures is to do away with such a reliance on monolithic RDBMSs and embrace alternative data storage mechanisms (such as SimpleDB as the AWS service or self managed mongoDB, CouchDB etc). When choosing where to store data on AWS, EBS is near the bottom of the list, right above getting monkeys to write it on paper and read it back to you.
Concluding that SSDs are faster/better/cheaper than EBS for MySql data is like concluding that bicycles are faster than cucumbers, they are in a different class and used for different purposes. The big concern is that this is not clear to the people trying to make use of EBS, particularly the MySql community.