Joining. Connecting. The path. IoT is the first step towards inter and intra connection as a state. Where all devices have a connected and many will also have a disconnected state. Data collected from some will be transient data, only used and consumed for a short time period. Other data will be less transient but not permanent have a TTL of weeks or months. Finally there will be the generated data that lives permanently.
Weather represents pure transient data. You only need it for the moment you are checking it and some small amount (12-48 hours) of trending. Sitting in front of the general store, or talking to your children or friends you can recall the hottest day ever back 20 years ago. But the relevance of weather data is pretty immediate. I need to know what the temperature is now, and what is potentially coming in the next 8 hours or so. Predictions of severe weather should go out as far as possible safely. Don’t warn me about a hurricane 12 months in advance, but a couple of days at least would be good notice.
The intriguing question is what about data that isn’t transient, you will in the end consult it more than right now. It applies not only to what you are doing but also what you may do next but doesn’t have to reside in the permanent memory space. A great example of this type of data is instructions for putting something together. You know permanently how to consume instructions but any one device you won’t specifically need to remember how to put it together. If you buy a new one you won’t remember how it goes together you will read the directions.
Long term or permanent data on the other hand presents an interesting question. The first is what do I do with that information. We keep information in our heads that is relevant to what we do. While I cannot spell worth a darn, never got spelling rules into my permeant memory I can tell you how to configure a connector between Microsoft Exchange and Lotus Domino. I haven’t actually done one in about two years, but I can still walk someone through it. Permanent memory is the stuff we use and reuse or recall frequently and it doesn’t get archived. It is interesting because many of us archive our earliest years of living before we learn how to manage our own memory. In the computer world we run into this in two formats. In my book Transitional Services I introduced a KM process I called DLM©. The concept focused on building, managing and acquiring IP related to answering questions. I talked about the KM concept of moving from the napkin to formal documentation quickly. Not all napkin creations are worthy of being IP. But all ideas are. In the DLM concept we create tiers of information. The top tier is managed by SME’s. My argument for services and IT organizations was that the SME’s should be rotational. I am a SME for 6-12 months and then I got back to the field and work or back to the IT shop and work for a time in that field (or another field). SME’s have two primary functions, they watch for interesting napkin ideas and they Shepard the existing IP as it is being used. IE people approach KM systems from the perspective of I have a problem, I need an answer. If they use IP you are managing but they modify that IP because of scenario variations, you now need to store both sets of managed IP the old, and the new IP. SME’s are also responsible for the lifecycle management of IP. Information doesn’t live forever. It has three distinct types (transient, not transient but not permeant, and permeant). The SME’s manage where in that pantheon the information exists right now. They also manage the variations within managed IP. Finally they spent a portion of their days fishing for napkins and great ideas.
I won’t go into the formal requirements around what is a SME. Just that personality is critical. If they believe they are the smartest person in the room your DLM system fails. SME’s have to be smart about the topic but willing to hear new ideas. That can be a VERY rare skillset.
So now we can break IoT into, the internet of devices that produce information, the decision tree as to where that information should be as far as the time based nature of the information and finally to complete this extremely long sentence who is going to consume the data.