Hi, I'm part of the 0MQ team in FastMQ, we're making a free low latency messaging library in C++ and Dhruva suggested it could be useful in Boost. The current stable 0MQ/1.0 is probably too complex to use in Boost but the forthcoming 0MQ/2.0 has a simpler socket-style API (sends messages rather than bytes) and runs with no external components. Dhruva is probably better placed than me to explain how 0MQ could help Boost users. One issue we've already discussed is licensing: currently 0MQ is LGPL, but we'd be prepared to change the license on 0MQ/2.0 to BSL (or dual-license it) so it can be included in Boost. Please excuse in advance my ignorance: I'm neither a developer nor user of 0MQ, but acting as community front-end. Cheers, Pieter Hintjens FastMQ
On 2009-07-28T14:06:55, Pieter Hintjens wrote:
I'm part of the 0MQ team in FastMQ, we're making a free low latency messaging library in C++ and Dhruva suggested it could be useful in Boost.
I think it would be great. /Allan -- Allan Wind Life Integrity, LLC <http://lifeintegrity.com>
I'm part of the 0MQ team in FastMQ, we're making a free low latency messaging library in C++ and Dhruva suggested it could be useful in Boost.
The current stable 0MQ/1.0 is probably too complex to use in Boost but the forthcoming 0MQ/2.0 has a simpler socket-style API (sends messages rather than bytes) and runs with no external components. Dhruva is probably better placed than me to explain how 0MQ could help Boost users.
One issue we've already discussed is licensing: currently 0MQ is LGPL, but we'd be prepared to change the license on 0MQ/2.0 to BSL (or dual-license it) so it can be included in Boost.
Please excuse in advance my ignorance: I'm neither a developer nor user of 0MQ, but acting as community front-end.
Very good idea! I'm interested in seeing this happen. A nice integration with Asio would be a cool thing to have. Regards Hartmut
Hartmut Kaiser wrote:
Very good idea! I'm interested in seeing this happen. A nice integration with Asio would be a cool thing to have.
OTOH that is probably somewhat different from the existing code - reading between the lines for the state of completion port support, anyway. This component would be much higher level than anything else in Boost, except perhaps quickbook and the preprocessor. Is it really desirable that it should be 'part of boost'? What is the point, compared to a project on Googlecode that uses Boost? I like messaging products and value choice, but this smacks of an attempt to compete with Apache qpid by buddying up to an existing project. Note that qpid already uses Boost, though once again its not an asio user. I think it should be considered, but with some caution. What is the intended approach to on-the-wire etc - is it going to be rewritten as an amqp 1.0 system, for example? James
Hello Pieter, Could you please have a look of "Channel", a message passing library i have been developing in boost, for a comparison with 0MQ? website: http://channel.sourceforge.net/ design doc: http://channel.sourceforge.net/boost_channel/libs/channel/doc/design.html download: http://www.boostpro.com/vault/index.php?PHPSESSID=9db4b8ef273029f908594e1689c4649e&direction=0&order=&directory=Distributed%20Computing Thanks Yigong On Tue, Jul 28, 2009 at 2:06 PM, Pieter Hintjens <ph@imatix.com> wrote:
Hi,
I'm part of the 0MQ team in FastMQ, we're making a free low latency messaging library in C++ and Dhruva suggested it could be useful in Boost.
participants (5)
-
Allan Wind -
Hartmut Kaiser -
James Mansion -
Pieter Hintjens -
Yigong Liu