On Tue, Dec 18, 2012 at 2:24 AM, Mark Hahn <m...@hahnca.com> wrote:
> I have an app where many servers will need to write to a store but
> only one server will be reading the store. The data will be somewhat
> transitory so an in-memory store might make sense.
Do you need the "store" to be distributed or a single instance
might work? What is the expected load (messages per second, message
size, total connections, etc.?) What do you use it for (general
pointers).
If you are interested in message queues there are a few options,
all with tradeoffs:
* http://www.rabbitmq.com/ -- implements AMQP, maybe the most
mature and complex (or flexible) solution;
* https://github.com/robey/kestrel (used by Twitter?) --
implements a memcache-based protocol (thus usable from almost any
programming language via a memcache client library);
* https://github.com/bitly/nsq (used by BitLy) -- implements a
custom memcache-like protocol;
* a dozen others (see
http://wiki.volution.ro/Mosaic/Notes/Solutions#Message_queues for some
that I've tried to "categorize");
But you could also use something based on a embedded database like
BerkeleyDB, LevelDB, etc. Maybe look at TokyoDB which has both a
memcache-based or HTTP-based access protocol.
Hope it helps,
Ciprian.
--
Job Board: http://jobs.nodejs.org/
Posting guidelines:
https://github.com/joyent/node/wiki/Mailing-List-Posting-Guidelines
You received this message because you are subscribed to the Google
Groups "nodejs" group.
To post to this group, send email to nodejs@googlegroups.com
To unsubscribe from this group, send email to
nodejs+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en
