Decision Management and the Internet of Things

I recently spent some time with the Bosch Software Innovations team (makers of Visual Rules, recently reviewed on the blog) and one of the topics of discussion was the Internet of Things – an area of research of great interest to Bosch as a whole. As Wikipedia puts it this refers to “uniquely identifiable objects (things) and their virtual representations in an Internet-like structure.” This begins by putting RFID tags or other sensors on physical things and connecting them, often wirelessly, so that you know where they are and what they are doing. The real-world, or at least the bit with sensors on, can then be tracked and a virtual representation manipulated using computers. Actuators and controllable devices increasingly allow these things to be manipulated in turn. Over time more of these devices can be expected to embed their own computing power and to be able to make their own decisions, often while collaborating with nearby objects as well or instead of communicating with some central network.

There is a clear link to Decision Management and to Decision Management Systems in the Internet of Things. First there is the fact that one of the main drivers for companies to build Decision Management Systems is a move to real time. Once a real-time response is required it is almost always impractical to have a person determine the appropriate response. This means a system must make a decision. When things are connected, using RFID or other sensors, they create a real-time flow of information about what is going on (or not going on) and Decision Management Systems will be required to determine the appropriate action to take.

For these kinds of systems you need business rules capabilities that can be easily integrated with event streams and that have low-latency deployments suitable for this kind of environment as well as analytic tools designed to execute against streaming data. Fortunately both these capabilities are actively being developed with a particular focus from the leading business rules vendors (including Bosch SI) to integrate event handling/state management with business rules.

The Internet of Things also lets organizations build Decision Management Systems that act on the real world not just consume data from it. Systems can now make decisions about what to do (based on inputs, locations of things etc) and then tell the things to take specific actions. We already see this with smart devices (such as smartphones in the hands of mobile customers) but we will see it more and more with physical equipment in factories or the outside world. One of Bosch’s clients, for instance, uses Visual Rules to process data streaming in from heavy equipment to determine that it should be sent for service or to give real-time advice to operators.

Bosch SI sees business rules as a key component of a software platform for the Internet of Things and I have to say I agree completely. The real world is even more changeable than the business world so agility is key for any system that interacts with the real world. The Internet of Things is also very heterogeneous with lots of manufacturers, lots of parts and models and constantly updated hardware to be coordinated. Being able to quickly and accurately add new rules to handle new pieces of equipment is critical to building a system that can be maintained easily as things change.

Of course building agile, analytic and adaptive Decision Management Systems also ensures that the huge volumes of data being produced by these things can be turned into analytics and consumed while also supporting the kind of test-and-learn experimentation that will be necessary to find optimal performance in such a complex environment.

In the long run the interesting thing is being able to push “intelligence” – decision-making – into the devices themselves. If you can give a device local decision making power, particularly if this decision-making can be constrained by overall policies and guidelines and can involve interactions with other things in the immediate vicinity has tremendous power in energy management, automating cities and factories and more. This is an area of active research for Bosch SI, making sure the rules can be pushed out into the network and embedded in these increasingly smart devices.

Related to this is the challenge for decision-making systems when consumers are involved – when consumers own some of the devices involved. Now the system must handle all multiple rule owners (power company, municipality, home owner) as well as data privacy. This is an area that Bosch is investigating and with which I have some personal experience from years ago and it certainly has its challenges – these systems don’t require more rules than existing systems but they have far more rule “editors”, requiring a much more embedded, seamless and idiot-proof rule editing mindset than a typical enterprise system.

Finally Bosch is doing research in some cool areas like mixing ERP systems and machine monitoring using business rules as well as using business rules for robot control. This latter area has some interesting potential for mass personalization, using rules to handle the variations each consumer wants while using robots to ensure they can be produced in scale.

The folks at Bosch SI have an Internet of Things blog too for those of you interested in the space.