[openamq-dev] Managed .NET client implementation?
Pieter Hintjens
ph at imatix.com
Wed Jun 13 09:54:57 CEST 2007
Karl,
> So, the ideal solution would be an OpenAMQ client.
> written completely in .NET. Are there any plans in this direction?
It just takes one .NET expert and a week or two... :-) We'd love to
make this. Do you know anyone who has .NET networking expertise?
> If there are, the second question is, which API to expose. There does
> not seem a long-standing tradition for a message queue API in the .NET
> world, so a direct translation of the WireAPI would be OK. However, the
> new WCF framework includes support for message-queueing, so one should
> probably also look into how the WireAPI could be plugged in.
Probably, the WireAPI is the essential API because it gives full
control over the protocol. But this can be layered with other APIs
that conform better to the language standards. We have the same issue
in Java, with JMS. Most existing messaging expects JMS, but to really
get the most from AMQP you need something like WireAPI.
What we expect is that in a few years people will all adopt something
like WireAPI for direct messaging work because it's simple and it
works the same in all languages. And on top of WireAPI we can make
all kinds of frameworks (e.g. APIs that deliver messages of specific
formats to applications).
-Pieter
More information about the openamq-dev
mailing list