Lecture 20: The Internet of Things

Like datacenters, the Internet of Things (IoT) is another use case for networking that has emerged over the last few years. And like datacenters, there is a history of previous technologies that were predecessors to IoT. Let’s break down the term IoT first. The term “thing” in IoT refers to any device that combines some mechanism to sense the physical environment, some amount of computation, and a means of communicating the sensed (and potentially) processed data to an interested entity. One example of a “thing” is an Internetconnected video camera at a road intersection. Here, the mechanism to sense the environment is a camera, the computation is likely an embedded processor integrated into the camera, and the camera has a WiFi or cellular network interface card to connect to the Internet. The term Internet in IoT refers to the fact that these “things” are all connected to the broader Internet. At least two attributes distinguish a “thing” from standard computers (e.g., servers, desktops, laptops, tablets, and even phones). First, “things” are made for a specific purpose and are typically constrained enough that they cannot run arbitrary applications—unlike a server/desktop/laptop that can run binaries downloaded from the Internet and a phone that can run mobile apps. Second, there is a sensory aspect to “things,” which allows them to integrate communication and computation with the external environment around them. Although, mobile phones today come equipped with an array of general-purpose sensors that sense mobility (e.g., accelerometers, gyroscopes, and GPS), the sensors on “things” tend to be far more specialized to the use case on hand (e.g., a temperature and humidity sensor for a soil and/or environment monitoring application, or a high resolution camera for a security application).