Skip to main content
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [m2m-iwg] AXEDA Comments on MQTT contribution to ECLIPSE PAHO & OASIS



On Mon, Jan 28, 2013 at 3:37 PM, Rick Bullotta <rick.bullotta@xxxxxxxxxxxxx> wrote:


Thanks for the feedback.  A few clarifications:

 

1)      When I refer to discovery, I am referring to discovery and browsing of subscribable topics and such


Aha, gotcha. Yes, good thought.
 

2)      We’ll have to agree to disagree on the critical importance of rich metadata.  Having built two platforms for industrial & M2M integration, I can attest to the incredible utility of this capability (and it is directly tied to discovery).  Though perhaps we are actually saying the same thing, just a question of whether it is part of the MQTT spec or some other spec.  If it becomes moved to another spec, I think it is important to set appropriate expectations as to what MQTT is and what it isn’t, as there is a great deal of confusing/misinformation in the market on that topic 

OK. I did wonder as I typed that, whether we are agreeing! :-)

My personal view is that MQTT is the transport, and the fact that it doesn't impose any structure on the data means that it's potentially very flexible. I agree that (as you and Matteo both say) understanding formats and metadata is critical (I did integration projects at IBM for 10 years myself, with MQTT and with WMQ and other messaging technologies, so I understand your view on this I think). I think that codifying data formats in the MQTT spec would significantly complicate it, though. 

Having the M2M IWG coordinate discussions and draw up specs for data, topic hierarchies for different industries etc is my understanding of the goal here, but I might be coming at things from the wrong angle.
 

LMK if there’s a public wiki and I’ll be happy to contribute some thoughts/content.


Well, there's the M2M section on the Eclipse wiki - probably best for metadata discussions etc.
For the protocol specifically, I already linked to the mqtt.org wiki where we've been pulling together thoughts and potential requirements for later revs. Would appreciate your input!

Andy
 

 

Cheers,

 

Rick

 

From: m2m-iwg-bounces@xxxxxxxxxxx [mailto:m2m-iwg-bounces@xxxxxxxxxxx] On Behalf Of andypiperuk@xxxxxxxxx
Sent: Monday, January 28, 2013 10:13 AM


To: m2m Industry Working Group
Subject: Re: [m2m-iwg] AXEDA Comments on MQTT contribution to ECLIPSE PAHO & OASIS

 

 

On Sat, Jan 26, 2013 at 8:49 PM, Rick Bullotta <rick.bullotta@xxxxxxxxxxxxx> wrote:

Well, someone at OASIS is trying to bring MQTT under their auspices, as I’ve received specific solicitations to that effect asking if we wanted to participate…

 

Yes, Scott has been working with OASIS I believe.

 

Also, I think a LOT more focus needs to be placed on some of the significant functional gaps in the current MQTT protocol,

 

Just before I jump into your list, I'm very ready to agree that we need to ensure that the 3.1 spec is properly clarified / non-ambiguous. That's what we've been doing over on the mqtt.org wiki at http://mqtt.org/wiki/doku.php/mqtt_protocol#clarifications as questions from implementers have arisen.

 

You'll see there is an Enhancements section there too, but I think that's beyond the scope of the initial OASIS discussion (I'm not specifically part of that process beyond my involvement with Paho and the M2M IWG).

and on capabilities that would greatly improve interoperability and integration.

Notably, these include:

-          Discovery capabilities

Agreed, although it's already possible to plug Bonjour / zeroconf in on top of a broker instance using e.g. avahi https://bugs.launchpad.net/mosquitto/+bug/902916 http://mqtt.org/wiki/doku.php/broker_auto-discovery so I am not convinced this needs to be part of the protocol. 

-          Stronger typing and metadata for payloads

-          Semantics for data, events, services, blob/file and stream content (all essential for a modern M2M platform)

These two I very strongly disagree with codifying as part of the protocol itself. I think one of the benefits of MQTT is that it is agnostic of the data layer, and only deals with the transport. Personally, I agree that we should think about data formats for M2M events etc, but that this is part of what the M2M IWG is intended to consider rather than something that's appropriate in the protocol - as I often say, it's a deliberately simple tool which you can use in a variety of ways. 

-          Standardizing a REST API model

Arlen Nipper was looking at this as part of the IWG activities.

 

-          Store-and-forward best practices/reference implementations for occasionally connected devices

Yep. I know Marco Carrier has requested that the Paho Java client expose additional methods for this. In general, more examples and good practices for these kind of scenarios would be extremely useful.

-          …and so on… 

Appreciate the thoughts above - we definitely need to get a lot more "written down somewhere" - I hope we can all share more experience on wikis and mailing lists as we go through this process.

 

Andy 

Those items would go a long way towards helping MQTT achieve its full potential as a general purpose M2M protocol, API, and standard.

 

Rick Bullotta

ThingWorx

 

 

From: m2m-iwg-bounces@xxxxxxxxxxx [mailto:m2m-iwg-bounces@xxxxxxxxxxx] On Behalf Of UOMo
Sent: Saturday, January 26, 2013 2:40 PM
To: m2m-iwg@xxxxxxxxxxx
Subject: Re: [m2m-iwg] AXEDA Comments on MQTT contribution to ECLIPSE PAHO & OASIS

 

All,

 

Thanks a lot for the update. I was about to ask about Paho as opposed to the Lua tools earlier, but had to re-subscribe due to that strange GMX mail bug with Eclipse.org.

At the moment, it doesn't look like there's a Paho or Java/C related M2M package under the Kepler release train, or did I miss something?

 

Should it still be part of those trains, how about Eclipse for Mobile Developers: http://www.eclipse.org/downloads/packages/eclipse-mobile-developers/junosr1

 

It contains both C and Java support, so it might be a good foundation for Paho.

 

Not sure, what's it with OASIS? It is mentioned a few time, but the message does not refer directly to anything by OASIS.

 

There is of course an OASIS standard, UnitsML often used as transfer format to the UCUM standard and codes: https://www.oasis-open.org/committees/tc_home.php?wg_abbrev=unitsml

 

Interesting, so while MQTT isn't an OASIS standard, it looks like AMQP is about to become one?;-)

 

Werner


_______________________________________________
m2m-iwg mailing list
m2m-iwg@xxxxxxxxxxx
http://dev.eclipse.org/mailman/listinfo/m2m-iwg



 

--
Andy Piper | Farnborough, Hampshire (UK)
blog: http://andypiper.co.uk   |   skype: andypiperuk
twitter: @andypiper  |  images: http://www.flickr.com/photos/andypiper


_______________________________________________
m2m-iwg mailing list
m2m-iwg@xxxxxxxxxxx
http://dev.eclipse.org/mailman/listinfo/m2m-iwg




--
Andy Piper | Farnborough, Hampshire (UK)
blog: http://andypiper.co.uk   |   skype: andypiperuk
twitter: @andypiper  |  images: http://www.flickr.com/photos/andypiper

Back to the top