It has been my experience so far that for many applications the higher speeds aren't justified with the associated costs. When I was working for an OEM, the extremely high-speed stuff was handled with a separate dedicated bus. The rest of the automation system didn't require the speed as much as the reliablity.
Not completely sure. Maybe others can weigh in. One Gig Ethernet is very commonly used for first tier consumer level gear. I suspect most installations, especially in control applications, are operating at 10 Gb/sec. Moving to the higher speeds for a typical installation just doesn't justify the extra cost at this point. But as the cost moves down, we know the need for speed will move up ... especially with the complexity of the plant-wide Ethernet networks growing.
Changes in industrial networking have always been "life in slow motion" especially with the amount of legacy systems running older proprietary protocols. It's not going to happen overnight. But except for simpler device-level networks, Industrial Ethernet protocols are the 800 lb gorilla in the room at this point when it comes to industrial control/factory enterprise networking and connectivity. The other networking technology to keep an eye on is WiFi.
I like this trend; I'm wholeheartedly behind it and reach for Ethernet/IP whenever possible in my projects.
(You can hear the BUT coming, can't you?)
This trend is riding on a venerable but misused component. Industrial networks should not be relying on RJ45 connectors. They're not robust and were never intended for the factory floor. I'd like to see a new standard industrial connector for Ethernet, and I'd like to see the companies involved for once agree quickly on it. M12 Code D would be nice, but I won't insist on it. But something better than what was originally designed for office and home use has to get onto the factory floor.
I guess I must be old. I look at a the combined Rockwell/Cisco proposed 'Campus netowrk, a single robust network infrastructure' where everything is connected as:
- cryptography - makes it harder to troubleshoot
- layers of firewall with many rules, which should be consistent but are REALLY easy to mess up
- selected services enabled 'by exception only' filtering unknown data at multiple points
- edge ports disabled by default, making troubleshooting more difficult
- a complex backup system required for the various switche configuration, with their various VLANS
- administration to keep the versions of all the switches synchronized and working together
I need data to flow in a timely manner to keep my mill running. If a problem arises, I need tools to figure out what is wrong, so the failed component (or errant rule) can be located and rectified.
Getting coherent log data out of a Cisco network, in my opinion, requires a lot (400 hours? per year) of training in Cisco classrooms. And personnel dedicated to network support.
Or I can put in dedicated IO networks (Profibus, Controlnet, Devicenet, Modbus Plus, and all the old ways) and I can troubleshoot them, one at a time, isolated.
Fixing an IO network is, compared to fixing an ethernet network, a piece of cake.
Right now, our over-worked but pretty well trained IT guys, can't tell me why it sometimes takes 2 minutes to open a 5K text file from our head office via Lotus Notes. I am quite technical on control systems, but not on ethernet - I already have a job.
With Rockwell saying that 60% of their business is Ethernet-based, we can only assume that there are a lot of people out there still using proprietary protocols. Any idea if we will ever see across-the-board Ethernet, Al?
Agreed; a great article. I think the most important element of the rise of Ethernet is that it's going to force vendors towards commonality, at least as far as intercommunications are concerned. OTOH, this won't eliminate vendor lock-in as much as one might suppose, because there are other proprietary elements involved in, for example, the tools chain used to program your PLCs. Writ large, we have a battle for control of the factory between the high-end (i.e., expensive), high-value vendors, and the low-cost, buy it online and do it yourself world.
Good point, Beth. That hadn't occurred to me. That certainly means that the plant technology is not foreign to IT.
Another trend that is helping matters is the formation of mixed groups that include members of both IT and control. These groups lead technology adoption and thus the needs of IT and control are both represented throughout the process of add new technology.
A new service lets engineers and orthopedic surgeons design and 3D print highly accurate, patient-specific, orthopedic medical implants made of metal -- without owning a 3D printer. Using free, downloadable software, users can import ASCII and binary .STL files, design the implant, and send an encrypted design file to a third-party manufacturer.
For industrial control applications, or even a simple assembly line, that machine can go almost 24/7 without a break. But what happens when the task is a little more complex? That’s where the “smart” machine would come in. The smart machine is one that has some simple (or complex in some cases) processing capability to be able to adapt to changing conditions. Such machines are suited for a host of applications, including automotive, aerospace, defense, medical, computers and electronics, telecommunications, consumer goods, and so on. This discussion will examine what’s possible with smart machines, and what tradeoffs need to be made to implement such a solution.