I recently attended a panel discussion where the three participants (usual suspects – easy to figure out who they are) proclaimed the future of “Wireless I/O” on the factory floor. Instead of cables running EtherNet/IP, PROFINET IO and Modbus TCP from your controllers to devices on your machine, you are going to use wireless and probably 5G.
Here’s why I don’t believe it, in no particular order:
COST – The “experts” told the audience that the cost of a wireless I/O point will soon be equal or less than the cost of a wired I/O point. I’d argue that they are slighting the true cost of the wireless link. When you factor in the preinstallation site costs, the expensive tools that you need to have on hand and the highly trained wireless technicians, wireless will for the foreseeable future be much more expensive than wired machine I/O solutions. The total cost is higher.
JITTER – Jitter (the variance in time delay for the arrival of a packet) on a wired network today is almost negligible given the typical wired infrastructure that we have in place at most factories. Jitter on a wireless 4G network is on the order of 4 milliseconds – much too large for many factory floor applications. I’ll concede that it is possible that the Ericssons of the world and other wireless companies may eventually have new technologies that reduce the wireless jitter in the future to be something close to what we can get on today’s wired networks.
SITE SURVEYS – Wireless networks are fragile; the higher the frequency, the more fragile than they are. I don’t believe that users will accept a networking solution that requires an AT&T, Verizon or other wireless carriers to conduct site surveys before a machine is constructed. I don’t believe that anyone is going to accept a system that might fail because a large metal scrap cart is moved. There will be applications for 5G, but I don’t believe I will ever see the majority of industrial I/O go wireless – 5G or any other wireless network.
TROUBLESHOOTING – Troubleshooting wireless networks requires special tools and special training, increasing the costs to the end-user. These skills will be very marketable, meaning that large end-users will be training somebody else’s workforce.
LEARNING CURVE – Every organization will have to go through an extensive learning curve to identify the best way to adapt wireless to their manufacturing environment. The time invested in that learning curve is time not invested in improving the reliability and quality of your processes through digitalization.
NEED FOR EXPERIENCED PERSONNEL – See Troubleshooting.
RESPONSE TIME – Wireless response times now approach 10msec. That’s adequate for many of today’s applications. Response time is not an issue in the future of wireless on the factory floor.
SECURITY – The security processes built into the wireless protocols we use today are adequate to prohibit attackers from using the network and a non-factor in the adoption of wireless. What is not apparent to me is how easy might it be to launch a denial of service (DOS) attack against a wireless I/O network in a plant. Could I, for example, drive a van by a factory or use a drone to fly over a factory and launch a DOS attack? It does not seem out of the realm of possibility that wireless networks could be subject to these sorts of attacks.
It’s going to be a long, long while before manufacturers will remove the wires and trust their I/O to a wireless system. It could be said that many people stated the same thing when wired Ethernet was introduced, but with wired Ethernet, there was a clear, measurable and significant performance increase to be gained. That isn’t true of wireless – wireless for the foreseeable future looks interesting but is nothing anyone is going to be in a rush to implement.