IIoT

My thought is that its all buzz words. its not gonna take over. MQTT itself is a nice concept but with bandwidth being cheaper and more readily available than it has ever been, it kind of makes it not worth investing in that infrastructure. Then no end devices support it which means you have are moving the traditional polling model of having the server do the polling to a cumbersome box called an edge of network gateway that basically moves the great tools most scada softwares have to create tags to this edge of network gateway.

I just dont see it being feasible long term. Not to mention that depending on your design, you may go from having updates every 5 minutes to updates every few seconds. Is 100 times more data in your database going to be useful? Maybe in some cases... but I think collecting a crapload of data just to do it is a waste of time, money, and space.

what it boils down to is after we are all dead and gone 70%+ of devices will still be using analog signals or one of the 10 flavors of modbus over a serial port.

We are a distributor of a lot of different industrial products and many of the sompanies are trying to come out with products to help people try to collect and or use the data they have.

Apple auto-correct has a different idea.

This topic is so big, there is no beginning or ending. But what I understand for me is simply that big companies will demand to have access to their machine data.
Take a look at transport. Are there still trucks where the home base does not know where they are? How long they still need to reach point B?

If a company has like 100 or more production machines all over the world, many of them much alike, they would not be interested in compare the productivity? Where is production smoother, where they have less quality issues, which parameters are in use, compare this and that.

The can build their own information database on predictive maintenance. When having access to many machines and parameters of the machines there may be created some simply rules... if this temperature is >> lim, it want take long untils this or that brakes. Lets change it now! We can shift production to another machine ...

But for that 1) data need to be accessible, 2) somebody has to analyse the data and create actions out of it.

IIot is about part 1) they companies have the desire to look into the data, only after some time part 2) will play the bigger role.

Some time in the future the data will be accessible (anyhow, no way to avoid it) and some smart companies will have big advantages over others where they can offer new services and new possibilities to there customers.

BR
BB

bits'bytes said:

If a company has like 100 or more production machines all over the world, many of them much alike, they would not be interested in compare the productivity? Where is production smoother, where they have less quality issues, which parameters are in use, compare this and that.

Click to expand...

That was done over 25 years ago via modem.

Peter Nachtwey said:

I looked up MQTT. It doesn't look very efficient.
ProfiNet was a huge help in my opinion.

Click to expand...

I think the benefits aren't when you compare it to something like Profinet; I think it fits better in telecontrol/RTU type situations. Based on my brief glances at it in the past, it is supposed to have security built in (so you don't need a vpn), be very tolerant of intermittent communications, and makes sense for applications that will have lots of peer to peer communications. I can't quite call it "decentralized", because as far as I know it does require a central server, but it is way closer than the traditional master/slave comms in most industrial protocols. The protocols don't use those terms any more, but the concepts are still present in the implementations, whether the protocols actually require them or not.

I don't know what the stack looks like (for device manufacturers) in comparison to existing protocols, though.

mk42 said:

I think the benefits aren't when you compare it to something like Profinet; I think it fits better in telecontrol/RTU type situations. Based on my brief glances at it in the past, it is supposed to have security built in (so you don't need a vpn), be very tolerant of intermittent communications, and makes sense for applications that will have lots of peer to peer communications. I can't quite call it "decentralized", because as far as I know it does require a central server, but it is way closer than the traditional master/slave comms in most industrial protocols. The protocols don't use those terms any more, but the concepts are still present in the implementations, whether the protocols actually require them or not.

I don't know what the stack looks like (for device manufacturers) in comparison to existing protocols, though.

Click to expand...

Im not aware of any major end device manufacturer that is putting mqtt in its devices. Since none do, you need an "Edge gateway". So in essence by having an "edge gateway", you are moving the typical stuff you do with a scada server, like setup tags, polling rates etc to this edge gateway. I would say in most instances this is alot more costly and also alot more cumbersome.

Now I do think that some portions of using MQTT makes sence, like making your existing tags available to many different applications... but I am not sure how many business type applications support this kind of protocol, although I know alot of the "big data" cloud providers like amazon, microsoft, etc do.

https://en.wikipedia.org/wiki/MQTT#Real-world_applications

As diat150 says I have not yet seen any manufacturer of industrial automation devices that incorporate MQTT.

For me the protocol of the future is OPC UA, which offers security and encryption and can work in both publish-subscriber and request-reply modes.

B&R is using MQTT to transport OPC UA data from the Edge-Controllers to the cloud as MQTT is completely unaware of what data it is transmitting.

In near future (maybe end of 2018) the operating system of the PLCs will also support MQTT embedded, meaning the PLC itself can connect to the cloud, for the moment only the 'Edge controllers' are capable of it (as long as I understood).

Wow -thanks to everyone who has contributed to this post! Apparently I'm not the only one who is both interested and skeptical about IIoT. I haven't figured out my presentation yet, but it is sure going to include some interesting viewpoints from this forum.

I joked around with my apple auto-correct but Bosch Rexroth has a product that includes multiple protocols including MQTT and is geared to be quick to startup. They call it the IoT Gateway:
https://www.boschrexroth.com/en/us/...lectric-drives-and-controls/iot-gateway/index

The link includes a description of a test case where they used the device in one of their manufacturing facilities.

Exor, traditionally thought of as an HMI company has devoted a lot of resources for IoT and Cloud development:
https://exorint.com/product/corvina_cloud_2/

Our company is having several sessions at our biennial event "Drive for Technology"
http://www.cvent.com/events/drive-f...summary-b69a5d91e5bb40218f0400365e0c1682.aspx

The event and technical sessions are free.

<disclaimer / I work for CMA/Flodyne/Hydradyne and we are distributor for the products mentioned above>

jraef said:

I just looked at some features in some new VFDs that will PREDICT when a cooling fan is about to fail, warning the operator of it (via Ethernet) so that PREVENTATIVE action can be taken BEFORE an unscheduled shutdown takes place. THAT is where the rubber starts to hit the road in my experience. I also just implemented a system wherein a pump motor run by a "smart overload" is reporting shaft kW, then in the SCADA system we are comparing that to flow out of the pump and if flow/PU of shaft power starts to deviate, we can predict when the motor bearings are going to need to be replaced. Again, TANGIBLE benefits. It's getting there, it's just going to take some edumacation of the workforce.

Click to expand...

Yours is a great example of how you don't need a Server and an expensive support contract with predictive software to determine certain things with a plant. It's down to whatever instrumentation and knowledge you have of the system and implement it.

I'm not an opponent or proponent here. I'm not getting into the should we/should we not discussion either. I have opinions on the subject, but I'm holding tight on all these "things of things", for now, and perhaps in the trenches while I wait and see if the mist of buzzwords dissolves into friend or foe?

I'm just here to provide some detailed and hopefully insightful information I had read not too long ago, for you Tom to ponder, and for others if they care to.

The following is from a forward thinker I do like to read, who like myself, does sense the inevitability of all this, whether we like it or not...

The Future of Automation in the Age of the Internet of Things by Walt Boyes

Working for a multinational, this is already blindingly on my horizon so I do like to keep abreast.

The later ICS Cyber-Security keynote by Joe Weiss is also an interesting read. Especially with regard to the lack of consideration, within the current industrial security standards, for cyber security at the architectural Level 0/1 i.e. device level. This most definitely will need to be addressed before disseminating such potentially critical data to the beyonds.

Anyway, I'm not getting into all this now, although it is hard not to express an opinion, or two.

Regards,
George

Interesting. At my company some engineers are involved in trying to implement IIOT but I am not involved in this project. I think it is being marketed to customers as a preventative maintenance strategy. Until reading this thread I didn't even realize how it can make you more vulnerable to hacking.

I made a similar post today
http://www.plctalk.net/qanda/showthread.php?t=119775

Tom Jenkins - I work with MQTT every day, and there are device's that have this feature built-in. I realize it's not the only part of IOT/IIOT but it does/will play a BIG part in this new world.

I also don't work in the "PLANT" environment, I work in the oilfields, where we use this protocol to communicate all of our data from the REMOTE units(which are in 7 states) back to the SCADA server(Ignition).

Now if you really what to understand MQTT you can talk with Mr. Arlen Nippers from Cirrus Link (He is the Co-Inventor) of MQTT. About 2 years ago he helped me to understand this new protocol, and that's why I am where I'm at today.

As always this is just my 2 cents worth.

Everyone have a GREAT day!!