La Internet de las Cosas (parte 2)



By Jim Walsh .

Let’s see a concrete and current example about an application based on the concepts of the Internet of Things. In this way we can know how these systems will work in the future. The vast majority of the examples provided for IoT revolve around a “smart coffee cup”, or household appliances with sensors. While these are valid examples, they focus too much on the “Things” aspect, which is only part of the IoT. The revolutionary of the Internet of Things are the systems that are intertwined from things. As time goes by, these advanced IoT applications will impact our lives as deeply as web 2.0 innovations did. As we will see below in our example, this is already happening today.

Uber is a mobile application that connects people with taxis and rental cars. Maybe it’s not the most obvious example for an IoT system, but I think it’s an excellent case of success, and it shows us the future of these platforms. In addition, it has the advantage of being a current and profitable system, which already has a concrete impact on the real world. I intentionally choose an example about which I do not have an internal knowledge, in this way I will not reveal any secret. This also means that I speculate with some of the details of the implementation, but I guess I’m not that far from reality.

This is the way Uber works today. First of all we must download the application from the operating system store, and then register with a credit card. The steps are the following.

1. When we need a taxi, we start the application from a mobile device.
2. The Uber app shows us our location on a map, along with the location of all the cars we can call. It also shows us the estimated time for the arrival of the vehicle. In my case, it was only two minutes.
3. While we look at the map, the location of nearby cars is updated in real time.
4. We can enter our destination and obtain an estimated price even before calling a car. In my case, the cost was from $ 23 to $ 31 for a 25-minute trip.
5. Once we decide to call a car, we press a button to confirm our location. At this point we can indicate our current coordinates or define a nearby location.
6. When the trip is confirmed we receive the driver’s name and a general description of the car, which includes the color and the license number. From this moment we will see the car on the map, approaching in real time on our mobile device. The reputation of the driver, in stars and comments, appears along with this information.
7. At this point the app shows us a counter and constantly tells us where the car we are hiring is. We also have an option to exchange messages with the driver, which can be useful if we change plans. When the car arrives we receive a notification on the phone.
8. Once inside the vehicle, the driver already has all our information, and receives step-by-step instructions to reach our destination.
9. When we arrive at our destination we simply appreciate the trip. No need to pay, leave a tip or even take the phone out of your pocket. The tip is included in the price, and the cost of the trip is automatically debited from the credit card. Then we received an email with a receipt for the operation.

And that’s it. I can summon a car with just a few clicks, and without using money or taking the phone out of my pocket. There are multiple wrinkles and refinements. For example, we can choose the type of car that interests us (taxi, limousine or personal car), which is reflected in the price we pay. It is also possible to qualify the car and the driver, share the fare and other tools for the passenger. Drivers can also qualify passengers, and receive recommendations on the most required places. All the drivers and passengers I spoke with agree that the system works very well. And besides, it’s really profitable. Uber earns its money through a commission of 20% on travel, which meant an estimated profit of 220 million in 2013, in a total of 1.1 billion transactions.

Uber is a good example of the Internet of Things in multiple aspects. In the first place, it is important to take into account the central role played by both the location of the car and the passenger, and the real-time nature of this information. The GPS chip, in conjunction with other positioning systems, allows you to identify the coordinates of any mobile device. While these sensors are contained within smartphones, they are the same as those used for the Internet of Things. In the case of Uber, the “Things” are the sensors of the mobile devices of passengers and drivers. The “Internet” in this case is the interconnection of each sensor, through the smartphone, with the brain and memory in the Uber cloud.

I think it is revealing that very few consider Uber as an example of the Internet of Things. It is a case in which the important thing is not the devices, but the people involved, that is, the passengers and the drivers. What happens is that mobile devices, together with their sensors, have become part of our daily life. We have a tendency to focus on people and human needs, such as transportation. In the future of the Internet of Things, our needs will always be at the center.

Continuing with our analysis of Uber, once the application is launched, the geolocation sensors of the devices send their information to a system that is hosted in the cloud. When we call a car, our device sends a message to the Uber cloud, indicating that the person X, in such coordinates, wants a car to take him to his destination.

When you receive an order of this type, the Uber cloud uses real-time analytics to determine the ideal vehicle for the passenger. I do not have the specific data of their algorithms, but they work combining geographical proximity with the estimated travel time, the reputation of the users involved and other factors. Most of this information is processed in an instant, using the freshest data from the mobile devices of passengers and drivers. A certain percentage of that information, such as rates, is processed according to a schedule, and not necessarily in real time.

The net result is that a car is quickly assigned to a passenger, approximately one tenth of a second. Because certain complex factors are calculated in advance, the system can use sophisticated metrics to improve the value of business decisions. In theory, this allows for better decision management at an impressive speed. Once the system obtains a result, it is put into effect. In the case of Uber, this takes the form of a notification to the driver, who notifies him of the passenger’s request.

Other applications of the Internet of Things tend to follow an architecture similar to the one we describe for Uber:

1. Orders, in addition to sensor and human information, are transmitted to a system in the cloud.
2. A quick decision subsystem processes observations and orders quickly.
3. This decision subsystem uses the context provided by the main system, which analyzes multiple sources of information.
4. Relying on the context, the decision system initiates an action, such as sending a notification.
5. Other complex actions can be initiated for analysis, as in the case of reputation or awards.

While this pattern adapts very well to the Internet of Things, it can also be used in multiple situations. Whenever an action must be contextualized in response to a flow of information, this approach will be very useful. We find that the elements of this IoT architecture are excellent for the fields of mobile advertising and computer security, for example.

The key to the Internet of Things is the ability to place information and orders in context, and then respond to them intelligently. When these observations come from sensors, the IoT label almost always applies. But the heart of any IoT architecture is the ability to intelligently respond to events through the autonomous decisions of a system. In the case of Uber, the heart of your business is the ability of your cloud to assign passengers and drivers wisely. When we talk about the Internet of Things, the important thing is not only the interconnected “Things”, but the intelligence generated based on the context of the information.