idea. The idea seems to be for Scraperwiki to focus on the ETL stuff and CKAN on the storage, versioning and registration. Last time I looked, the ETL layer was based around building out VMs (virtualbox IIRC) that might add a lot of weight to a CKAN install.