Amazon SimpleDB vs Amazon RDS

SimpleDB can only scale by sharding, has 10 GB data size limit per table, and query performance is parallel to record count (eg: poor if you have 1 million records). And google's datastore is slower than simpledb. Cassandra is much more scalable, high traffic sites began to use it, there is nothing better for free if you need high write rates with massive data. cassandra survey

If your read/write ratio is something like %90 for read and %10 for write, then terracotta or infinispan with postgres is a better fit. There some free clustering options for postgresql but none of them matured (mostly prototypes).

Another option is sharding. Hiberntae and NHibernate has sharding support. You can use them with postgres or mysql but you loose joins.

Regards

Amazon RDS is pure MySQL, accessible by your app the same way as any other MySQL database; the web services interface to RDS is purely for creation, deletion, and modification of the DB instances, not the DB data. From their FAQ:

Q: How do I access my running DB Instance?

Once your DB Instance is available, you can retrieve its endpoint via the DescribeDBInstance API. Using this endpoint you can construct the connection string required to connect directly with your DB Instance using your favorite database tool or programming language. In order to allow network requests to your running DB Instance, you will need to authorize access. For a detailed explanation of how to construct your connection string and get started, please refer to our Getting Started Guide.

This is the part of the Getting Started Guide you need -- it explains how to get the hostname of your new instance so you can connect to it, authorize the instance for access from the client, and then connect using the MySQL command-line client (as an example):

$ rds-describe-db-instances --headers
$ rds-authorize-db-security-group-ingress default --cidr-ip 192.0.2.0/30 --headers
$ mysql -h myinstance.crwjauxgijdf.us-east-1.rds.amazonaws.com -P 3306 -u mymasteruser -p