It would be useful to have a JDBC/Datasource protocol do perform the autodiscovery of other nodes.
For example when used together to Infinispan which might use a database as cache store it would be practical to be able to reuse the same datasource for this purpose, so that setting up a service requires a single set of credentials.
In case of deployments on EC2 so it could be possible to use S3 for both discovery and cache store, or RDS for both discovery and cache store (or mixed in any form).