Hi > For my purposes, I would let the database manage conflicts. Each application scope object would > be tied to the database, a change made on a specific application server would first update the > database. For example, the information from a table containing label/value pairs for product > categories would be stored in application scope on each server at system startup. If a change > were made to the product categories, the change would first made to the database and then a change > notification would be broadcast to all servers in the cluster, each server would then reload the > information from the database. >
I'm not sure how often your data MAY change, but looking at the above explanation, why don't you simply use small piece of code for checking data stored in the database? You can add a column with timestamps, which can be checked, for example, every five minutes. I know it's not too elegant solution, but I'm sure it saves you the overhead, added by JMS-based solution. Sergiusz --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
