On 5/30/05, Barry Beattie <[EMAIL PROTECTED]> wrote:
> so what you're saying is that holding or caching of data across the cluster 
> should be avoided because the mechanics of refreshing each instance is the 
> tricky bit.

No, if all you're doing is caching then each server instance should
manage its own cache anyway and you wouldn't need to propagate data
across the cluster.

> if an instance gets an update, it knows where the other instances' 
> webservices are and "pushes" the new data to them. since each instance know 
> where it's "mates" are any one of them can update the whole lot, yes?

Yes, each instance would have a list of other instances and when an
instance gets an update that needs to be propagated, it would invoke
the web service on each of its "mates".

> That seems to be a bit of work to set up. I take it you'd only do this 
> sparingly then...?

Yes, I'd say that in reality this is a rare case. You do not need to
propagate cached data across a cluster - although you may need to
propagate a cache invalidation across a cluster. In general, if you
have data that needs to be managed across a cluster, a database is
usually a more effective way to do it that to try to handle it via
inter-instance communication.
-- 
Sean A Corfield -- http://corfield.org/
Team Fusebox -- http://fusebox.org/
Got Gmail? -- I have 50, yes 50, invites to give away!

"If you're not annoying somebody, you're not really alive."
-- Margaret Atwood

---
You are currently subscribed to cfaussie as: [email protected]
To unsubscribe send a blank email to [EMAIL PROTECTED]
Aussie Macromedia Developers: http://lists.daemon.com.au/

Reply via email to