This year, Interop's live production network takes on the challenges of networking the Internet of Things.
When we hear the phrase "Internet of Things," we tend to think of small devices, such as household appliances -- thermostats, security cameras, and electric car chargers.
These types of objects have been on the Internet for a long time. The first Internet toasters were running as part of the Interop show network in 1989. In 1990 and 1991 there were Lego robots (to insert and remove the bread from the toasters), model railroads, weather stations, talking bears (giving weather reports), juke boxes, and vending machines. (Simon Hackett demonstrates the Internet toaster at Interop in 1990 below).
Like the part of an iceberg that is visible above the water's surface, the task-doing part of Internet of Things is but a small piece of the whole. The invisible, but much more significant, pieces of Internet of Things are the parts that do management and control.
If you look at the Internet of Things with a wide-angle lens that encompasses management and control in addition to the doing-the-job functionality, it brings you into the realm of "process control systems." These kinds of systems are common in industrial settings. But with IoT they will become part of our daily lives.
Process control systems involve many pieces:
• Pieces that are sensors
• Pieces that are actuators (things that do something)
• Management pieces that sit between the sensors and actuators.
Management pieces have the job of figuring out what to actuate, when to actuate it, and how long to keep actuating. An example of a simple household process control system is a thermostat.
This device includes:
• A thermometer (a sensor)
• A clock and user-control panel (management)
• Some circuits (actuators) that turn on heaters and air conditioners.
This is not a complex system, and we used similar technology when we hooked toasters and talking bears to the Internet back in 1989. But that was 1989; it is now 2016 and there are some new aspects to IoT.
One new aspect is that the "simple system" may now include additional sensors and actuators and, most particularly, management somewhere in "the cloud" of the Internet. For instance, your IoT thermostat may chat with the electrical utility to negotiate the best time and prices to heat or cool your house.
Another new aspect is that these systems may operate largely without human attendants or human intervention. And in some cases they are designed so that human intervention is difficult. (For instance, IoT-based home systems try hard to assure that burglars can't take control.) The lack of human intervention is actively debated by those designing self-driving vehicles: Do human controls -- such as steering wheels and brake pedals -- increase or decrease the reliability and safety of the overall system?
When viewed in this light, IoT starts to take on a different complexion. It becomes less about the things themselves and more about the control and monitoring matrix that supports those things.
Issues raised include:
• Security: As our dependence on Internet attached appliances increases, so does the risk of harm should those come under hostile control. The Stuxnet virus that attacked uranium enrichment centrifuges is a well-known example of Internet of Things security vulnerabilities.
• Data mining/privacy: IoT creates new opportunities to observe the activities of people and enterprises. This is a privacy problem. In addition, there is a wide open question regarding the ownership and control of information generated by IoT devices. Since IoT devices will often be part of an international matrix of systems, there may be issues with trans-border data movement and collection.
• Bias: As decision making moves into IoT, there will be opportunities for suppliers to influence those decisions. For example, consider an IoT system that you can use to buy groceries and have them delivered to your home. Which store will it buy from? If you ask for a generic product, such as milk, which brand will it select? We can easily foresee that a business can be made selling the rights to influence those choices.
• Maintenance, finding, and disposal of old devices: Things wear out, are replaced, and sometimes are upgraded. Replacement parts may disrupt the operation of the parts that were not replaced. That risk should not be underestimated -- new parts are often not tested against older pieces, and older parts are often not designed with an eye towards robust service under changed conditions. People who make repairs may often substitute parts that are incorrect. For example, as we move toward self-driving automobiles we may see collisions and injuries caused by improper or negligent replacement of sensors.
• Lock-in: Once a set of compatible IoT devices are deployed, there is an understandable hesitancy to make changes in the future that might undercut the original investment. This will create a lock-in effect that will keep old or expensive IoT technologies alive and operating well after they have become obsolete.
• Signal pollution: Many IoT devices use unlicensed radio spectrum. That spectrum is already getting tight. A deployed IoT system might begin to go awry if a new user of that spectrum comes along. But this issue of competing and disruptive use of communication channels is not limited to unlicensed radio. We are likely to see IoT problems arise as competition increases for optical, infrared, and audio signal channels.
• Growing regulatory frameworks at technical, national, and international levels: The US government is already reviewing whether new laws or regulations are needed for IoT. For example, the Department of Commerce has an open inquiry inviting interested parties to weigh in and provide input and concerns regarding the government's potential role in regulating IoT.
• Deployment of IoT will result in some damages. Who will be liable? Can we even find the proximate cause of a malfunction or flaw? And what will be the legal remedies to cure, or at least compensate, for harms incurred?
As we enter the new world of IoT, we must ask the right questions and put in the right safeguards to assure that in each instance we have considered control, monitoring, security, data mining/privacy, maintenance, lock-in, regulatory compliance and liability.
This week at Interop Las Vegas, the InteropNet Demo Lab team is examining these issues and the technologies being introduced to alleviate them. Messaging middleware, interoperability, standardization, and improved security are the keys to growth. The team is demonstrating options live on the Expo floor and hosting educational sessions on these topics and more.