We are currently thinking about the following architectural scenario:

• A number of application servers that form an Infinispan clustered cache
• A "backend" Infinispan cluster with file based cache store
• The backend cluster is the cache store for the app server's cache

It would be nice if the backend cache could inform the app server's cache about cache entry invalidations.

This could be used when you want to use Infinispan clusters in your "branch offices" and have a "central" Infinispan clusters that receives all data updates and informs the "branch" clusters just about invalidations.