Bernd Hofner
unread,Apr 13, 2012, 10:57:31 AM4/13/12Sign in to reply to author
Sign in to forward
You do not have permission to delete messages in this group
Either email addresses are anonymous for this group or you need the view member email addresses permission to view the original message
to Akka User List
Hello,
I have about 25 million objects, that can be uniquely identified by a
10 digit number.
I need to collect and process a bunch of information for each of these
objects.
After everything is processed, I need to retrieve the final state for
each of my 25 Mio objects (sorting them into "good" and "bad").
Is it a "good idea" to create an Actor for each of these objects?
I need to look up each Actor several times (let's say 10 times).
Or, to put in other words: I have about 10 independend data streams,
potentially sorted by my 10 digit object id and need to feed the
relevant portions of these streams into the corresponding data object
- which will change its state accordingly.
Will the lookup for each Actor (via its ActorPath with my unique ID as
a name) be efficient?
Does it make sense to place the ActorReferences into a separate
datastructure (index-like) to find them quicker?
Or would it be a better idea to keep my objects passive and just
create a few workers that lookup the "right" data object (let's say
from a Hazelcast Map) when a message comes in, modify it with the
incoming message and place it back in the Map?
Best regards,
Bernd