Re: LookupCache, ServiceDiscoveryListener and discard guarantees

Gregg Wonderly wrote:

> Okay, so you made me actually remind myself of the SDM API and the
> LookupCache. So the version of "lookup" that returns an array of
 > "ServiceItem" actually does provide the type of functionality I
 > described. I had forgotten about this API being available.  I had
 > not used SDM, early on, because of the fact that it would not
 > populate if callbacks were not available.  So, that's really
> the only valid reason why I'm doing what I'm doing.

Does that mean you would like to see a second event model for Jini
that doesn't require callbacks or the set-up of a MailBox service that
you have to poll your-self.

'Pull under the hood' style based events might scale less, but besides
the advantages as set out in "Jini accross the Firewall" they also
simplify the configuration of security if that is a concern and as such
I prefer that model for any service of which I expect it to be reachable
beyond the LAN. Actually due to the way the JSC API is written pull
versus push events is for the service implementation rather irrelevant,
the client decides which model it utilizes and in the future for JSC 1.0
they can even switch for a given event registration on the fly.

I didn't include ServiceRegistrarX or the pull style event model as part
of the work I envision for River in the JCM10 presentation as I can't
recall a very warm welcome in the past and since I've become an even
stronger proponent of pull versus push. So if nobody pulls it in, I'm
not going to push ;-) Another advantage of pull is that the pullers have
to do some work too, but enough 'hints' for now.
--
Mark

--------------------------------------------------------------------------
Getting Started:     http://www.jini.org/wiki/Category:Getting_Started
Community Web Site:  http://jini.org
jini-users Archive:  http://archives.java.sun.com/archives/jini-users.html
Unsubscribing:       email "signoff JINI-USERS"  to [email protected]