+1 386-243-9402 MON – FRI : 09:00 AM – 05:00 PM

The MIT Technology Review this week had a nice story about how GE is promoting the idea of an “industrial Internet”. This is a really good idea. A number of us at IBM have been talking about “millions of businesses, billions of people, and trillions of things” connected to the Internet for quite a few years. A short story I wrote 5 1/2 years ago about the topic appears below. We are really at the beginning of the industrial Internet even though it has been around for quite a few years, at least in concept. Someday I will count all the light switches, sensors, computers, audio and video devices, etc. that are connected to my home network – which in turn is connected to the Internet.
GE sees a big opportunity in collecting data from the locomotives and aircraft engines it produces so they can provide better service to their customers. IBM began doing this several decades ago, albeit without the Internet. IBM connected customer’s mainframe computers to a telephone line and monitored what was going on — not the customer’s data but the performance of the mainframe. I remember one day when a customer engineer stopped by the desk of the CIO at one of my customers and said that he would be bringing a new power supply for the mainframe in the next day or so. The CIO said there were no problems with the computer and asked why the power supply was going to be replaced. The customer engineer said the power supply was fine now but it was projected to fail the following week. Analysis of data from the mainframe collected over the phone connection showed  a deterioration in the condition of the power supply and projected when it would fail. This was very basic compared to what can be done today. I am sure GE will be using advanced analytics capabilities to examine huge amounts of data from their products to determine what components might be near failure and to provide data to their customers that may result in improved efficiency.
In the healthcare arena, the potential for telemonitoring as an aid to the treatment of chronic illness has huge potential for reduced costs from hospital readmissions and improved outcomes for patients. Rather than monitor a critically ill patient in intensive care every hour or even every five minutes pales compared to the potential to use sensors to monitor vital factors thousands of times per minute and predict impending problems in time for clinicians to intervene. The healthcare section of patrickWeb has other stories about the role of technology in healthcare and many more to come.

Reflection – written July 20, 1997

This is a view of the future from one fellow traveler of the planet. It has to do with a universally connected world which I believe is emerging at a breakneck pace. This is not a new thought but I may have a new twist on it, especially when it comes to the kind of opportunities such a connected world may present.
What I see happening is the emergence of a global local area network. A network built on open standards. A network soon to have, perhaps, accessibility by a billion people. People everywhere having access to this global area network, enabling every business, every institution, every government to think of their local area network as a way to reach all of their constituencies. In addition to large numbers of people being connected I believe we will have even larger numbers of things connected. At the turn of the century maybe a billion people and a trillion things. Internet addresses everywhere. I’m not talking here about e-mail. I’m talking about the ability to address things. We know that every PC has a TCP/IP address. Soon, everything will. Your phone. Your pager. Your car. Can you imagine a vending machine sending a message to headquarters saying “I’m out of 7-Up”? Can you imagine your KitchenAid dishwasher sending a message to the server in your basement, sending a message in turn back to headquarters saying “Please send a repairman. I need a new impeller blade”? Can you imagine your car running a little Java applet sending you an e-mail saying “It’s time for an oil change”?
Some people think that we will never have enough TCP/IP addresses to enable all these things to be connected since there is talk of running out of addresses already. However, the next generation of TCP/IP, called IPng or IPv6, is just around the corner and it has exponentially expanded addressing potential. One estimate puts the limit at just under 1,600 IP addresses per square meter of the planet. That assumes a less than optimal allocation of the addresses. Clever addressing may result in a limit closer to one million million addresses per square meter! It surely appears that addresses will not be a limitation. In fact various schemes already exist to connect and address large numbers of things without using tcp/ip at all. A company called Echelon has enable millions of devices around the world to be connected using a simple protocol they call LONWorks.
The Result
One Internet is what is evolving. A universally-connected world. And this universally-connected world will result in the natural evolution to the new medium. A new medium where everything and everybody can communicate. Portions of this single Internet are cordoned off behind firewalls to protect against unwanted intrusion: This is what we call the “intranet.” Portions of this single Internet are very high speed, very private connections between Company A and Company B, or Department A and Department B. Some people call these connections “extranets.” But it’s all one network and people and devices will be able to communicate if they want to. The result will be a whole new world of data and applications. Huge numbers of tiny cameras, sensors, monitors, communications relays, etc. will be connected and will create new data. Some of this data will reside on a very local level providing control feedback to manufacturing processes or home controls. However, a lot of data will be “rolled up” from these micro networks to LANs and WANs which will feed the data to databases for recording and analysis. This will provide the basis for the creation of applications that will have a dramatic impact on information technology as we know it, not replacing it, but extending it.
Related information
Paul Saffo at the Institute for the Future has written a paper called “Sensors: the Next Wave of Infotech Innovation” that presents an expanded and related view of this area.