First, let me say that I am not technically knowledgeable enough to make detailed recommendations about industrial communications and fieldbuses for any user's specific application. Also, since I run a company that is completely focused on industrial fieldbuses and networks, I do have a bias. With that in mind, I would like to comment on what is going on in the factory and process automation control markets. I think we are hurting not only ourselves but also the potential gains for our customers.
In the past eight to 10 years excellent open standard fieldbuses for field I/O devices have been developed and released (not without some struggles, of course). Fieldbuses have been supported by literally hundreds of companies that have developed thousands of products that can and have been used to improve the quality, cost, and efficiency of users' machines, processes, and factories. Open networks are providing benefits to people who are forward-thinking enough to consider and use them.
So how is this situation getting messed up? By letting the "Ethernet Monster" get in the way of taking advantage of these open networks and fieldbuses. This is happening in two ways. One, we are hurting ourselves by confusing Ethernet with open fieldbuses in what is possible and profitable. And two, we are again in the midst of the struggles that go on in the standardization of any technology in the industrial control world.
If I see one more study, article, or marketing piece that puts Ethernet together with DeviceNet, Profibus DP, AS-interface and Foundation fieldbus H1, I will quit golfing. They don't belong together! Ethernet will not do what they do cheaper, easier, or better. And the device-level fieldbuses can't do what Ethernet-based networks do effectively. Programs using Ethernet are great for data collection, data transmission, and monitoring. Ethernet can give the corporate IT person the data they need. Just keep it out of the control world! On the other hand, fieldbuses are great for collecting and managing the I/O data the machine and process control systems need. Most advanced automation architectures need both data and control networks to run well.
The other reality is that Ethernet needs to be "ruggedized" to survive on the plant floor. Using office type products and putting "thick clothes" around them will not work. Ethernet components that are used on the plant floor must be designed from scratch to work in a factory. If this isn't done Ethernet will have an adoption rate similar to early vision systems that had problems working reliably on the factory floor.
There is no question that Ethernet will continue to make inroads into the industrial market. It will not be cheap or easy, and it will not replace device level networks. Ethernet network design is very complex, and because it will be vital to operating the plant, it must be designed correctly. Most importantly, Ethernet will not replace other lower level networks. On the contrary, Ethernet will enable and advance the adoption of fieldbuses. An Ethernet backbone is needed to fully use the large amounts of data and diagnostics that fieldbuses provide. So let's do it right!