Hi,
The pattern states that
"The ValueListHandler session bean caches ValueList to avoid repeated, unnecessary execution of the query."
How is the caching performed so that multiple clients can share the same ValueList? And how can the server update the ValueList while a client is iterating over it?
Thanks.
A session bean may not be you ideal solution instead you may want to have a ValueListHandler object. If this is a web based app you may store that in the application context. You can also make that listhandler receive JMS asynchronous notifications when the values have changed so that it can refresh its data by possibly making a copy and changing the reference quickly.
Dan
I would point you to Tangosol's Coherence product (http://www.tangosol.com/coherence.jsp). This is a Clustered Cache DataStore implemented with a standard java.util.Map interface. It includes support for replicated or distributed datastores as well as implementing a locking scheme and the javaBean event model.
rmisek
Hi,
> A session bean may not be you ideal solution instead
> you may want to have a ValueListHandler object. If
> this is a web based app you may store that in the
> application context. You can also make that
> listhandler receive JMS asynchronous notifications
> when the values have changed so that it can refresh
> its data by possibly making a copy and changing the
> reference quickly.
Do you mean to say that this pattern is independent of the EJB ?
If for the web applcation I have around 1000 records dont you feel there will be 1000 object space allocated all the time.But definetly this will improve the efficiency of the search fires as it will be getting the data from the objects in the memory.So i can say I am achieving this quick web search at the cost of the memory allocations.
But what about the very big search where there will be 100000 + records ? Don't you feel the system will become very slow in that case.
regards
Vicky
