Hmmmm, what would be the point of doing a local invalidation when this node inserts the data? Would it invalidate the data it's inserting itself? I'm rejecting this for now. Feel free to send a PR with any proposed changes. Btw, the Infinispan cache provider now lives in the Infinispan source tree and you can submit issues [here](https://issues.jboss.org/projects/ISPN/summary) choosing Hibernate Cache as component.