On 2011-01-25 14.21, Niclas Hedhman wrote:
I think you made that sound a LOT easier than it in reality is, if you are going to allow for arbitrary failures of any underlying component.
Not really. For the "replication of data", which is what I talked about, it should be pretty much that simple.
A lot of scalability infra is not around 'replication of data', but fan out to enough compute (or even memory) power. For typical web requests you are fine, since the client waits and it can all be within a transaction of that request, but once you introduce fire-forget semantics, things quickly get very tricky... That is my day job right now :-)
That is another, also important, problem. Unrelated to EventSourcing, though.
/Rickard _______________________________________________ qi4j-dev mailing list [email protected] http://lists.ops4j.org/mailman/listinfo/qi4j-dev
