IoT / IIoT stuff

IoT / IIoT stuff

[slos]

Hello there,

The next big thing is marketed as Internet of Things and Industrial Internet of Things and I'd like "Ada for Automation" to jump in the train.

I have found that Mister Per Sandberg has done some work on the protocols used in this domain:
https://github.com/persan/zeromq-Ada
https://github.com/persan/mosquitto-ada

It seems also that Data Distribution Service (DDS) could also pretend. DDS is a standard supported by the Object Management Group (OMG), providing a data-centric publish-subscribe programming model for distributed systems.
Some implementations may provide an Ada binding.
I have found that most are open source and commercially supported.

Do some of you have some thoughts and experiences to share on these subjects ?

Best Regards,
Stéphane
http://slo-ist.fr/ada4autom

Re: IoT / IIoT stuff

[Dmitry A. Kazakov]

On 2016-01-21 14:25, slos wrote:

> The next big thing is marketed as Internet of Things and Industrial
> Internet of Things and I'd like "Ada for Automation" to jump in the train.

There is a lot of hype running all the time. Another one up right now is 
Industry 4.0. There were dozens (or hundreds if counted per application 
area) before. E.g. ASAM (Association for Standardization of Automation 
and Measuring Systems). I don't remember all of them and their ugly 
acronyms.

Give it 5 years, maybe we will forget this one too...

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

Re: IoT / IIoT stuff

[slos]

Le jeudi 21 janvier 2016 15:36:54 UTC+1, Dmitry A. Kazakov a écrit :
> On 2016-01-21 14:25, slos wrote:
> 
> > The next big thing is marketed as Internet of Things and Industrial
> > Internet of Things and I'd like "Ada for Automation" to jump in the train.
> 
> There is a lot of hype running all the time. Another one up right now is 
> Industry 4.0. There were dozens (or hundreds if counted per application 
> area) before. E.g. ASAM (Association for Standardization of Automation 
> and Measuring Systems). I don't remember all of them and their ugly 
> acronyms.
> 
> Give it 5 years, maybe we will forget this one too...
> 
> -- 
> Regards,
> Dmitry A. Kazakov
> http://www.dmitry-kazakov.de

Sure there is a lot of marketing hype around this.
My first thought was similar to yours but I have changed my mind because, even if the technology isn't as new as it is claimed, it is the way applications and services will be developed which is new, somewhere in the cloud with high number of connected devices.
Let's see.

Stéphane LOS
http://slo-ist.fr/ada4autom

Re: IoT / IIoT stuff

[Dmitry A. Kazakov]

On 2016-01-22 16:24, slos wrote:

> My first thought was similar to yours but I have changed my mind
> because, even if the technology isn't as new as it is claimed, it is
> the way applications and services will be developed which is new,
> somewhere in the cloud with high number of connected devices.

This is not how automation applications are designed and are required to 
be, as they know exactly where devices are and which sensors and 
actuators they have.

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

Re: IoT / IIoT stuff

[tmoran]

> This is not how automation applications are designed and are required to
> be, as they know exactly where devices are and which sensors and
> actuators they have.
  Lots of cheap sensors and actuators implies designing for unreliable, but
multiply redundant, sets of devices.  eg "Homeowner detected leaving work
at usual time, then detected at end of driveway at expected time, but no
detection of house door being opened.  What's the probability he's
actually home and the house door sensor failed, vs he went to the store
and the end-of-driveway detection was spurious, and what should we do?"
  But the use of Ada is orthogonal to that.

Re: IoT / IIoT stuff

[Dennis Lee Bieber]

On Fri, 22 Jan 2016 18:59:28 +0000 (UTC), tmoran@acm.org declaimed the
following:

>> This is not how automation applications are designed and are required to
>> be, as they know exactly where devices are and which sensors and
>> actuators they have.
>  Lots of cheap sensors and actuators implies designing for unreliable, but
>multiply redundant, sets of devices.  eg "Homeowner detected leaving work
>at usual time, then detected at end of driveway at expected time, but no
>detection of house door being opened.  What's the probability he's
>actually home and the house door sensor failed, vs he went to the store
>and the end-of-driveway detection was spurious, and what should we do?"

	... vs he had a heart attack trudging through the snow from the
driveway to the house door... <G>
-- 
	Wulfraed                 Dennis Lee Bieber         AF6VN
    wlfraed@ix.netcom.com    HTTP://wlfraed.home.netcom.com/

Re: IoT / IIoT stuff

[Dmitry A. Kazakov]

On 2016-01-22 19:59, tmoran@acm.org wrote:
>> This is not how automation applications are designed and are required to
>> be, as they know exactly where devices are and which sensors and
>> actuators they have.
>    Lots of cheap sensors and actuators implies designing for unreliable, but
> multiply redundant, sets of devices.

How can you build a redundant system without knowing what is there, 
being redundant or not?

The point is that any automation system always deals with some physical 
real life process. The inputs and outputs are mapped onto some physical 
entities. A sensor is there to measure a concrete entity of the process. 
You can use many sensors to measure the same entity in a redundant 
manner or you can have subsystems dormant when some measurements are 
unavailable, but you always know what you measure and what you control. 
A sensor in a "cloud" without reference to the process at hand has no use.

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

Re: IoT / IIoT stuff

[Georg Bauhaus]

Dmitry A. Kazakov <mailbox@dmitry-kazakov.de> wrote:
> On 2016-01-22 16:24, slos wrote:

>> the way applications and services will be developed which is new,
>> somewhere in the cloud with high number of connected devices.
> 
> This is not how automation applications are designed and are required to 
> be, as they know exactly where devices are and which sensors and 
> actuators they have.

The programs know exactly how to extract 
relevant, statistical information from the sensors,
and they are guided by BigData.

Another example is the determination of a price
per customer, IT-enhanced customer that is, in a shop.
Where is she, where was she before, what type of customer might she be
having been there and bought that?

Re: IoT / IIoT stuff

[slos]

Le vendredi 22 janvier 2016 16:44:08 UTC+1, Dmitry A. Kazakov a écrit :
> On 2016-01-22 16:24, slos wrote:
> 
> > My first thought was similar to yours but I have changed my mind
> > because, even if the technology isn't as new as it is claimed, it is
> > the way applications and services will be developed which is new,
> > somewhere in the cloud with high number of connected devices.
> 
> This is not how automation applications are designed and are required to 
> be, as they know exactly where devices are and which sensors and 
> actuators they have.
> 
> -- 
> Regards,
> Dmitry A. Kazakov
> http://www.dmitry-kazakov.de

IoT, IIOT, Industry 4.0 and comparable initiatives from Japan or US of which I don't remember the name at the moment deal with optimizing processes, energy efficiency, consumption of water or any raw material, work hours... and creating services to help you buy, knowing where you are and what you can't resist to.
Of course this has not same requirements as industrial automation but the data used in industrial automation can have some real interest for optimizing processes.

Re: IoT / IIoT stuff

[Maciej Sobczak]

> The next big thing is marketed as Internet of Things and Industrial Internet of Things

There is nothing "next" in this, especially if you are looking for communication solutions. These things have been solved in '80s and since then only variations of selected features have been developed.

If you are looking for something that is Ada-friendly and that can be also used with small embedded devices, then have a look at YAMI4:

http://www.inspirel.com/yami4/

You will also certainly find PolyORB being mentioned in the Ada ecosystem, but I'm not sure if it fits the IoT hype due to the memory and computing resources found on many network-enabled embedded devices. The "new" part of IoT is "T" and many of the existing solutions just do not fit.

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

Re: IoT / IIoT stuff

[slos]

Le vendredi 22 janvier 2016 12:48:24 UTC+1, Maciej Sobczak a écrit :
> > The next big thing is marketed as Internet of Things and Industrial Internet of Things
> 
>> There is nothing "next" in this, especially if you are looking for communication solutions. These things have been solved in '80s and since then only variations of selected features have been developed.
> 

For sure nothing has changed after Babbage, Turing or Von Neumann... ;-)
In the 80's TCP/IP was already there but not Internet as it is now with everyone having a PC or a Smartphone. With many challenges still to solve.
 
> If you are looking for something that is Ada-friendly and that can be also used with small embedded devices, then have a look at YAMI4:
> 
> http://www.inspirel.com/yami4/
> 

Thank you for the link. I have found interesting things there.

> You will also certainly find PolyORB being mentioned in the Ada ecosystem, but I'm not sure if it fits the IoT hype due to the memory and computing resources found on many network-enabled embedded devices. The "new" part of IoT is "T" and many of the existing solutions just do not fit.

Not all IoT devices need to be constrained. Gateways can exist and abstract a network of legacy devices for example.

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

Re: IoT / IIoT stuff

[Per Sandberg]

Well from my point of view the the important aspect is that we agree on 
a reasonable number of protocols on different levels in in the 
communication stacks and the protocols and standards (official or 
defecto) may be: IP, TCP, UDP, Spread, 0Mq, MQTT, AMQP,  Protobuff, 
Json, Corba, DDS,  or GumboJumbo.

I actually don't care which one and on what level we are talking , but I 
do care that the community agree on communication standards that are 
language neutral an includes the full semantics of the data flows from 
application down to the lowest level of transport.

The 0MQ bindings and the MQTT bindings was something I needed for some 
small projects (hobby and commercial).

You are quite right concerning DDS there is an Ada binding to the RTI 
implementation of DDS where I also was "involved".

I think that the next step would be to get a protobuffer implementation out.

/Persan

Den 2016-01-21 kl. 14:25, skrev slos:
> Some implementations may provide an Ada binding.
> I have found that most are open source and commercially supported.
>
> Do some of you have some thoughts and experiences to share on these subjects ?

Re: IoT / IIoT stuff

[slos]

Le mercredi 27 janvier 2016 06:18:42 UTC+1, Per Sandberg a écrit :
> Well from my point of view the the important aspect is that we agree on 
> a reasonable number of protocols on different levels in in the 
> communication stacks and the protocols and standards (official or 
> defecto) may be: IP, TCP, UDP, Spread, 0Mq, MQTT, AMQP,  Protobuff, 
> Json, Corba, DDS,  or GumboJumbo.
> 
> I actually don't care which one and on what level we are talking , but I 
> do care that the community agree on communication standards that are 
> language neutral an includes the full semantics of the data flows from 
> application down to the lowest level of transport.
> 
> The 0MQ bindings and the MQTT bindings was something I needed for some 
> small projects (hobby and commercial).
> 
> You are quite right concerning DDS there is an Ada binding to the RTI 
> implementation of DDS where I also was "involved".
> 
> I think that the next step would be to get a protobuffer implementation out.
> 
> /Persan
> 
> Den 2016-01-21 kl. 14:25, skrev slos:
> > Some implementations may provide an Ada binding.
> > I have found that most are open source and commercially supported.
> >
> > Do some of you have some thoughts and experiences to share on these subjects ?

Hello Persan,

Thanks for providing the mosquitto-ada binding.

I've been trying it on my Debian Jessie and I've succeeded to have it working although provided Mosquitto libs are 1.3.4-2 where you state 1.4 is needed.
I've had some trouble with the relocate scenario but things are fine with the static one.
I don't know yet if it has something to do with the old GPS provided or some stupid error on my side.
Anyway, it looks like it may not be long before "Ada for Automation" floods the cloud !

Best Regards,
Stéphane LOS
http://slo-ist.fr/ada4autom

Re: IoT / IIoT stuff

[slos]

Le mercredi 27 janvier 2016 17:31:43 UTC+1, slos a écrit :
> Le mercredi 27 janvier 2016 06:18:42 UTC+1, Per Sandberg a écrit :
> > Well from my point of view the the important aspect is that we agree on 
> > a reasonable number of protocols on different levels in in the 
> > communication stacks and the protocols and standards (official or 
> > defecto) may be: IP, TCP, UDP, Spread, 0Mq, MQTT, AMQP,  Protobuff, 
> > Json, Corba, DDS,  or GumboJumbo.
> > 
> > I actually don't care which one and on what level we are talking , but I 
> > do care that the community agree on communication standards that are 
> > language neutral an includes the full semantics of the data flows from 
> > application down to the lowest level of transport.
> > 
> > The 0MQ bindings and the MQTT bindings was something I needed for some 
> > small projects (hobby and commercial).
> > 
> > You are quite right concerning DDS there is an Ada binding to the RTI 
> > implementation of DDS where I also was "involved".
> > 
> > I think that the next step would be to get a protobuffer implementation out.
> > 
> > /Persan
> > 
> > Den 2016-01-21 kl. 14:25, skrev slos:
> > > Some implementations may provide an Ada binding.
> > > I have found that most are open source and commercially supported.
> > >
> > > Do some of you have some thoughts and experiences to share on these subjects ?
> 
> Hello Persan,
> 
> Thanks for providing the mosquitto-ada binding.
> 
> I've been trying it on my Debian Jessie and I've succeeded to have it working although provided Mosquitto libs are 1.3.4-2 where you state 1.4 is needed.
> I've had some trouble with the relocate scenario but things are fine with the static one.
> I don't know yet if it has something to do with the old GPS provided or some stupid error on my side.
> Anyway, it looks like it may not be long before "Ada for Automation" floods the cloud !
> 
> Best Regards,
> Stéphane LOS
> http://slo-ist.fr/ada4autom

Hello Persan,

On my Debian Sid machine where Mosquitto is at version 1.4.7.1 it works too.
And the relocatable option is fine too. I'd just omitted to install the library or run the tests with :
LD_LIBRARY_PATH=../../lib ./mosquitto-tests-main

Thanks !

Best Regards,
Stéphane LOS
http://slo-ist.fr/ada4autom

Re: IoT / IIoT stuff

[tonyg]

I commissioned some work on a ada binding to the openzwave interface. As to which I will be releasing as open source in the next few weeks. Its not 100 per cent complete and needs some examples, it'll soon be on github.