Hi, As Jukka already wrote, the Global Data Store feature should solve this problem. Certain operations like getProperty and getLength will be very fast, and no temporary copies are made.
> ...but i've been wondering for a while whether the need for clustering > really must imply storing everything in a database. I don't think this is a requirement: why not use a failsafe file system. Or storing large objects in regular file systems, but in multiple places. > why wouldn't a remote simple DAV store work? That would be a third implementation of the DataStore interface. I think file system storage and database storage will be more popular, but WebDAV and Amazon S3 storage are possible as well. Thomas
