MQTT native Ada implementation

MQTT native Ada implementation

[Dmitry A. Kazakov]

The incoming version of Simple Components will provide a native (not 
bindings) implementation of the MQTT 3.1.1 stack and on top of it a full 
MQTT messages broker.

In order to test the implementation I would be thankful for MQTT test 
use cases.

-- 
Regards,
Dmitry A. Kazakov
http://www.dmitry-kazakov.de

Re: MQTT native Ada implementation

[Per Sandberg]

Great
Then i cold discontinue the bindings
	https://github.com/persan/mosquitto-ada
and pass them into oblivion.
Do you implement support for the TLS transport as well ?
/Per


Den 2016-03-24 kl. 20:15, skrev Dmitry A. Kazakov:
> The incoming version of Simple Components will provide a native (not
> bindings) implementation of the MQTT 3.1.1 stack and on top of it a full
> MQTT messages broker.
>
> In order to test the implementation I would be thankful for MQTT test
> use cases.
>

Re: MQTT native Ada implementation

[Dmitry A. Kazakov]

On 2016-03-26 09:03, Per Sandberg wrote:

> Do you implement support for the TLS transport as well ?

Yes. The implementation is driven by a connections server object. The 
connections server may run several protocols simultaneously. The server 
itself has a number of implementations, one of them is a secure server 
that uses GNUTLS as the SSL/TLS layer. That should do it, I think.

MQTT with TLS must be quite slow with the QoS level 2, which deploys a 
lot of small packets sent here and there. And in general, MQTT design is 
not very exciting, but it was easy to implement, so I gave it a try.

-- 
Regards,
Dmitry A. Kazakov
http://www.dmitry-kazakov.de

Re: MQTT native Ada implementation

[Maciej Sobczak]

> Great

Yes.

> Then i cold discontinue the bindings
> 	https://github.com/persan/mosquitto-ada
> and pass them into oblivion.

No.

They have different licenses and this alone means that they have potentially different targets. Also, they have different underlying implementations, which means that they have potentially different sets of bugs and run-time parameters. There can be lots of reasons to use one variant over another.

Also, abandoned Ada libraries is a sign to the software community that Ada is giving up.

Also, competition is good. Why not compare them with some benchmark or something?

-- 
Maciej Sobczak * http://www.inspirel.com