Re: Weird Caching Questions

[Jon Dart]

Andreas Maschke wrote: Hi Dear All,   being a newbee to EJB I hope the following questions don't bother you: :-) I have a static class XYZData with a static Vector containing some data. The Vector acts like a cache. If I request some record from the class, it first searches the Vector. If there is no matching record it is loaded from the Database and appended to the list.  Now I have a stateless session bean which imports this class. There is a method which calls the static method of the class XYZData to request records. The questions are:  How many instances of the class XYZData (and the cached data) are created by an Application Server if it creates multiple instances of the session bean?  Is this a valid design to implement a (read-only) cache in EJB? Are there better ways? (Using "read-only entity beans" seems to be very vendor-specific.) You could certainly do what you're describing. You will get one copy of the static data cache per VM. If you run one app server instance in a single VM, you get one copy. If you run a cluster of app servers, though, each one will have its own cache. However, this isn't a very common programming paradigm. Keep in mind that your database server is going to cache frequently read data anyway. If you have a nice fast database server and effective connection pooling (so you're not having the overhead of reconnecting), you may find that pulling this frequently-accessed data out of the DB is not a great deal less effective than pulling it out of your cache. Also you may have a problem with your architecture re updates. E.g. if some person or process ever updates your DB tables directly, you will have a stale cache in your server. Some app servers can automatically maintain a memory cache for you, and can even manage a shared cache that is accessed across multiple server instances in a cluster (e.g. GemStone, Persistence). But this is not a standard feature.