IoT: From Cars to Stores
We need to think bigger for IoT to get big.
If there's any truth behind the rampant IoT publicity wave, IoT has to do stuff that's not already done. Sadly, most of the current IoT successes being touted are reprises of things we've had for years. Connected cars are an example; we've had in-vehicle cellular services, integrated with collision detection and vehicle health, for a long time. Remember OnStar? Despite the pedestrian (no pun intended) connected car apps today, the fact is that cars probably do figure into IoT's evolution.
Vehicle navigation is largely a cost-based hierarchical routing process not unlike finding a path in a network. The success of an auto GPS depends on its ability to provide proper weights for the routes under consideration. The most logical evolution of the connected car would be to have IoT sensor data on road conditions available to guide route selection.
But just thinking about this mission for a minute reveals issues with the traditional approach of sending IoT data to the auto GPS. Suppose you're driving from New York to Los Angeles. There are a lot of possible roads for a GPS to consider, lots of route updates and traffic status to process inside a GPS even without IoT data. With IoT, you might need a massive car computer and not just an Internet connection.
A better way, given that you're in a connected car, would be to send the request for a route to a cloud-based application that could track road conditions everywhere and run the necessary calculations in seconds. That process could also record the routes it had given out, and send an alert to vehicles if something like an accident changed the picture. Such an approach could work with any device being carried or built into the vehicle. You simply send the appropriate route, and perhaps even maps, as needed.
What's interesting about this is that it opens up another possible connected-car application. The largest number of auto-miles driven today is associated with getting to and from work, but the second-largest is linked to shopping or going to a restaurant. Some auto GPSs will let you "find" retail outlets, but the data isn't always up-to-date and complete, and in most cases it's not really what the driver wants. I want a good meal, in a BYO restaurant, with a nice view. Try telling your GPS that.
This is also a route-cost application. You can rate restaurants by their suitability and the difficulty in getting to them, then pick the "best" and guide the driver to it. If you're looking for goods instead of eats, you could rate the stores based on price and reputation, using product review data already available. "Take me to the best deal on high-rated 4k flat-screen TVs."
It gets even better. Suppose you want to meet some friends for a drink or meal. Your connected car can contact the proper cloud service and get a selected meeting spot considering the sum of the locations of all those involved, without having to spend long minutes working out who can get to where.
Then there are preferences. Suppose you had a bad experience with a retailer or restaurant, and so you'd not consider them again? Suppose you had a large group and you wanted to sit together, so you needed to link with a reservation system to nail down a table? Suppose you or someone else wanted to avoid driving winding roads after dark, so you had to rule out destinations that required that? You can see that there are plenty of perfectly reasonable constraints on retail selection that could demand a lot of information-gathering and processing.
IoT at Enterprise Connect
Enterprise Connect Regular Rate ends Friday, March 4!
Register for an entire event pass or Tuesday through Thursday conference pass to save an additional $200 off the regular rate. This special discount, available using the code NJPOST, represents a total savings of $500 off the onsite price! Plus, register three or more attendees from your company for even bigger savings.
Also note that this information doesn't have to be IoT-related. Retail pricing, table reservations, and personal preferences are traditional databases or APIs. These examples, then, demonstrate that IoT's utility rises if you can couple in data beyond IoT. They also illustrate that IoT could be helpful the other way around, feeding information to applications like reservation systems. If a restaurant knows that a bunch of people are heading its way because they've gotten directions and routes, the restaurant could prepare better for the group. Retailers and manufacturers could do "test buyings" that use the driver/shopper's tools just to see if there are hot spots of product demand that will be impacted by scheduled road construction, and try to deal with the issue by offering discounts or something of the like.
None of this works if you think of IoT as your car talking to a red light, any more than it would be validated by your refrigerator talking with your toilet. Sending or exposing vast clouds of data to an auto navigation system is similarly short of the mark. Simple machine state won't enrich our knowledge enough to be truly compelling. Simple applications suitable for in-car analysis don't take us far beyond OnStar. We need to think bigger for IoT to get big.
That's the bottom line, I think. Too much of our IoT focus is on connecting sensors and cars, things we've done for years. The future has to be driven by futuristic things if it's going to be different from the present, or past. IoT could be huge if our ideas match its potential.
Follow Tom Nolle on Google+!
Tom Nolle on Google+