combining reliable request-reply with multicast destination

[bandar]

hi room,

As document say that reliable request-reply channel will need reply message from destination.
if destination is a multicast processor which hold, say, 3 processor.
should all those processor send reply message to the channel, or only 1 processor should do that ?

Regads,
bandirsen

[Martin Krasser]

Hi bandirsen,

Am 13.05.13 06:21, schrieb bandar:

If n multicast targets reply, n replies will be sent back to the channel. A relieble request-reply channel however only processes a single reply, others are ignored. If you'd rather like to aggregate replies from n multicast targets before sending a reply, you'll need to implement that logic yourself (maybe Akka already offers a scatter-gather pattern out of the box that can be used in combination with the multicast processor). There's also a related ticket - let me know if it fits your needs and I'll schedule it for 0.6.

Regads,
bandirsen

--
You received this message because you are subscribed to the Google Groups "Eventsourced User List" group.
To unsubscribe from this group and stop receiving emails from it, send an email to eventsourced...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

-- 
Martin Krasser

blog:    http://krasserm.blogspot.com
code:    http://github.com/krasserm
twitter: http://twitter.com/mrt1nz

[bandar]

hi martin,

thanks for your reply,
how come i miss to check issue list page, shame on me.
my current scenario fit with current behaviour, so it wont be a problem.
however, after reading the ticket discussion, i think it would be nice to have that feature which support "n-synchronous processor inside a multicast processor" scenario

anyway, congrats for 0.5.0 release

regards,
bandirsen

[Martin Krasser]

Am 13.05.13 08:56, schrieb bandar:

hi martin,

thanks for your reply,
how come i miss to check issue list page, shame on me.

no worries :)

my current scenario fit with current behaviour, so it wont be a problem.
however, after reading the ticket discussion, i think it would be nice to have that feature which support "n-synchronous processor inside a multicast processor" scenario

I'm not sure what you mean with "n-synchronous processor inside a multicast processor". Are you referring to the second scenario in this comment?

[bandar]

hi martin,


***  I'm not sure what you mean with "n-synchronous processor inside a multicast processor". Are you referring to the second scenario in this comment?

yes, I'm referring second scenario:

sender processor(1) --> channel(1) --> multicast-processor(contains n-processor)
the sender processor need to wait confirmation reply from 2 or more processors inside a multicast-processor before continuing his work,

or maybe it would be better if we can mix it
for example a multicast-processor contains 5 processors,
2 processors receiving a message from a channel and sending back a confirmation reply
while other 3 just receiving the message and never sending back a confirmation reply


regards,
bandirsen