Hi all,
I am wondering what may happen with the curve handcheck mechanism in the
following situation. I am making some assessments of how I think it
works, with questions marks waiting for confirmation or corrections :
Say we have a server somewhere in the web, and two clients behind the
These sound like good ideas. I think we lack documentation on the
event interface, and it could be worth wrapping in CZMQ too.
I think formalizing the internal protocol would be good, and using a
default inproc endpoint as we do with ZAP could make it simpler.
On Fri, Sep 27, 2013 at 8:57 PM,
Formalizing the protocol would be good, but a default endpoint would make it
difficult to monitor 2 different sockets from 2 different places in code
would it not? Not sure if anyone uses this use case but I could imagine a
large zeromq based application might want to.
Ric.
From: Pieter
Right, for PUSH and DEALER you can use this approach.
On Fri, Sep 27, 2013 at 8:59 PM, Artem Vysochyn
artem.vysoc...@gmail.com wrote:
I was wrong about detecting sent/no-sent messages on PUB socket
using poller and POLLOUT events. I have muddle up this with PUSH,
because for PUSH this
On Sat, Sep 28, 2013 at 10:24 AM, Laurent Alebarde l.aleba...@free.fr wrote:
From the server, both clients have the same IP, but different ports
allocated by the company's router ?
Sure.
So the identities of the clients are different for the server's ROUTER
socket and there is no collision
I am using ZEROMQ for distributed messaging application. Need to connect client
(DEALER socket) to multiple servers (ROUTER socket on server side). What are
my options on CLIENT side ?
1. Create DEALER socket on client side for each server endpoint (ROUTER
socket).
2. Create
If you connect a DEALER to multiple endpoints it will distribute messages to
all of its connections. I think this is a round robin distribution.
If you have some specific logic where you need to choose which server to talk
to, you would need a router socket so that you can specify the address.
I imagine I am missing something basic here, I can see where the certs are
being generated in the examples but I am having a hard time tracking down a
few more things...
1. In the Ironhouse example the integrity of the client is verified, but I
am missing where the integrity of the server is
Hi guys,
I messing around with JNI and libzmq and I was wondering if there was
a way to avoid the memcpy?
JNIEXPORT
jobject JNICALL
Java_org_zeromq_jni_ZMQ_zmq_1recv__JI (JNIEnv *env, jclass c, jlong
socket, jint flags)
{
zmq_msg_t msg;
zmq_msg_init (msg);
zmq_recvmsg ((void *)
I've managed to fix the problem and now am able to use boost::serialization
to pass serialized data via zeromq. I've posted the test program code below
in case it might be of use to someone else trying to do the same in
future..There is a class declaration, a server program and a client program.
zcert and zauth are in czmq, not libzmq, and pyzmq doesn't wrap czmq. But
all of these things can be implemented with socket options, etc., which are
up to date with master in pyzmq. Once some of the auth APIs stabilize, I
will work out what helpful utilities belong in pyzmq (e.g. easy ZAP
Thanks MinRK, as always your work is fantastic. These higher level
functions being directly in libzmq will be crutial to making the bindings
smooth here and I am excited to see the work he moving forward.
I will keep code diving and looking for places to jump in and help.
On Sat, Sep 28, 2013 at
Thanks for the feedback. I will look at lbbroker examples.
In case I can't use the router broker socket approach in my use case, what do
you think about creating multiple DEALER sockets on client side - one for each
server endpoint (ROUTER socket)? Is it recommended by zeromq best practices?
13 matches
Mail list logo