> The JMS docs have an excellent description of the problem, and why
> it is so difficult to solve.
>
> I know that lateral cache synchronization is an incredibly difficult thing
> to pull off.
I've been thinking about this issue also. Seems like the JCS distributed
cache implementation could be modified to utilize versioning similar to that
used in optimistic concurrency control, i.e., in a lateral cache an object
instance would never overwrite an object instance with a later version.
Likewise, you need to utilize OCC between the various VMs and the database.
I think the combination of OCC between caches and between VMs and the
database would solve data integrity issues.
Performance, though, would be an issue if there was high-contention in your
application.
-Richard
--
To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]>
For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
