Pre-Combo AMI

A single 0.7.2 AMI
A single 0.7.4 AMI
A single 0.7.5 AMI
A single 0.7.x auto-updating AMI
- Where the cassandra package would install on launch
A single 0.8.x auto-updating AMI

Combo AMI

1.0
- A DataStax Brisk/0.8 AMI
2.0
- 0.7.x, 0.8.x, and Brisk on one AMI
2.1
- DataStax Enterprise and DataStax Community
  - Specifically for the DSE/C 1.0 release
  - Added CQL Shell
  - Added Demos
2.2
- DataStax Enterprise 2.0 and DataStax Community 1.0
  - Specifically for the DSE 2.0 release
  - Completely rewritten codebase for easy reading and understanding
  - Solr assignment and integration
  - Cluster names with spaces now allowed
  - New tokentool that supports multiple datacenter spacing
  - Removed all short arguments (-v) and only long arguments are accepted (--version)
  - --release allows for specific DSE package releases
  - hadoop.tmp.dir is now set to be on the RAID0 device, if applicable
Future (as of March 24, 2012)
- Planned DataStax Enterprise 1.0 support in AMI 2.2
  - Within a week
  - Made possible via the --release <version> argument
- Planned DataStax Community 2.0 support on AMI 2.2
  - After release
  - Will also allow for --release flag for DSC

Additional features are being incorporated via the CassandraLauncher as found here: https://github.com/joaquincasares/cassandralauncher

By using the CassandraLauncher, certain aspects are simpler to deploy from a central, non-EC2 location.

The list include:

DataStax username and password memory
AWS API key memory
Instance size memory
Input validation
Automated RSA fingerprint checking
Automatic separate known_hosts file
Pre-built SSH strings
Automatic OpsCenter agent installation
Passwordless SSH around the cluster
datastax_ssh tool
- Allows for SSH commands to easily be run across an entire cluster
nodelist file
- An uploaded file onto the cluster for later, possible, use
Modified hosts file
- Allows for easy jumping between machines, e.g. c0, c1, a0, a1, s0, ...