Hi,

Cache invalidation based on a cheap query result, like "select max(timestamp)", before an optimistic transaction or a cached reporting query can be quite useful technique sometimes : in clustered environement or when the same DB is concurrently updated by non-hibernate applications.

It this possible with Hibernate today? Is it the responsibility of the application to do the job, or do you think that Hibernate could provide such option in future versions? Is it the second level cache who should do it?

Is there a possibility to implemment such logic as a "plug-in" for hibernate?

Of course, it's not necessairily "select max(timestamp)", it can be any cheap query specified by application developers for cache invalidation in distributed environement and/or with non-hibernate DB clients.

Thanks
Vitaliy Shevchuk