Cunning way to scale out data services

[Dan]

Hi again fellow kettle peeps

So; Been exploring the kettle data service architecture and it's pretty nice.  (Previously called the kettle thin jdbc driver or something)

However - How can you scale that out. If you build a platform that may have 100's or 1000's of data services, can you push these out on carte somehow?

Is it possible to have a single connection point for the jdbc url, but the master then distributes you to one of the cluster slaves?

(Maybe this is possible, maybe not? If not I'll jira it!)

Thanks,

Dan