IoT 101: Connectivity

This is the first in a series of tutorials we’re calling IoT 101. Our goal is to provide an overview of the fundamentals with links you can follow for detailed information. Everything we’ll cover applies to developers of IoT devices and apps, whether or not you’re working with ARTIK. Let’s open with the most basic decision you’ll need to make about your Internet of Things: How to connect your things.

If you want to start an argument, just ask, “What’s the best radio network?” on an Internet of Things forum. The right answer, as for most design choices, depends on what you need your IoT to do and how much you’re prepared to spend to do it. In this tutorial I’ll explain the tradeoffs among eight radio network options common in IoT networks: four variations of WiFi®, two types of Bluetooth™, ZigBee and Thread. All options are supported in Samsung ARTIK chips; the information below applies whether you’re using ARTIK chips or not.

This tutorial covers two of the most basic properties of IoT networks:

  • How they transmit a signal through a radio channel (the physical, or PHY layer).
  • How they control who is allowed to transmit a signal through the channel (the medium access control, or MAC layer.)

I’ll cover routing, security, error correction, and other network considerations in a future tutorial.

Radio choice is a compromise among the bandwidth of the channel, the range of the radio, and power consumption.

The best radio for you depends on how much data you need to move across your IoT network, how long you need batteries to last, and how much infrastructure you can afford to interact with your network. In radio terms, you need to know the bandwidth, power consumption, and range of your radios.

The above chart gives an overview of the options covered in this tutorial. Bandwidth is along the horizontal axis, range is vertical, and the size of each circle indicates power consumed while the radio is active.

Local Area Networks

Use cases: WiFi connections work best for video and other high-bandwidth data streams, including nodes connecting a network of IoT devices to a larger network. If you need WiFi nodes to operate on batteries, pay careful attention to software techniques to minimize power consumption.

What we call WiFi today began 25 years ago when the first working group met to formalize the IEEE 802.11 radio standards. The concept of radio networks operating without government-issued licenses was still young; the first sensor networks I worked on required site licenses for every radio we used. With the opening of the airwaves it fell on industry and private engineers to devise standards and systems to prevent interference. The first standard released by the 802.11 working group supported three modulation options: direct sequence spread spectrum, frequency hopping spread spectrum, and infrared. Of those options, only direct sequence persists today.

Three variations of WiFi are supported in ARTIK chips:

  • 802.11b radios support up to 11 Mbps of usable throughput and have been available since 2000. They use the same direct sequence spread spectrum modulation supported in the initial (1997) specification.
  • 802.11g radios double usable throughput to 22 Mbps and came on the market in 2003. They use a different modulation technique called Orthogonal Frequency-Domain Multiplexing (OFDM) to achieve the speed boost.

Tip: Most access points today support 802.11b and 802.11g radios, but throughput for an 802.11g client can be reduced in the presence of 802.11b.


  • 802.11n radios surge usable throughput to 54 – 600 Mbps and came on the market in 2009. They use sophisticated signal processing and a special type of antenna called Multiple Input, Multiple Output (MIMO) to achieve these speeds. The specifics are complicated, but the concept is simple. One signal travels in a straight line from sender to receiver, and another bounces off a wall on the way so it arrives slightly later. Each signal carries as much data as the underlying modulation technique allows; combine the data streams and you boost the throughput between sender and receiver.

Tip: Sending, receiving, and processing multiple signals is probably not something you want to do in an IoT node running on batteries, but it’s a great option for a concentrator node relaying multiple video feeds from remote sensors.

Personal Area Networks

Use cases: Personal Area Networks work best when IoT nodes need to interact directly with each other, for example in home automation networks. Bluetooth radios are most popular with sensor nodes, while ZigBee and Thread are popular for power meters and output nodes where security is a concern.

The first mobile phone with Bluetooth shipped ten years after that first meeting of the 802.11 study group, ushering in an era of personal area networks. Conceived as cable replacement technology, Bluetooth and related radios have always been designed with networking in mind. Personal area networks organize into small independent groups called piconets. The distance between piconet nodes and the maximum number of nodes varies by technology.

  • Classic Bluetooth supports 1Mbps throughput and allows optional 2 Mbps and 3 Mbps connections. Most classic Bluetooth radios have a range of approximately 10 meters and a limit of 8 nodes in a piconet.
  • Bluetooth Low Energy radios support 1 Mbps usable throughput. Range up to 100 meters is possible in part because they use a modulation technique that can tolerate more interference but carry less data. There is no technical 8-node limit as in classic Bluetooth, but in practice piconets in excess of 10 or 15 active nodes are unusual.
  • ZigBee and Thread use radios based on IEEE 802.15.4 and are optimized for low bandwidth, extremely low power operation. Many devices can operate off coin batteries for months or even years. Most devices operate in the same 2.4 GHz band common in WiFi networks, but the standard also supports operation at 915 MHz. Networks with thousands of nodes are possible.

Explore the IoT 101 series: Networks

and Sensors

About the author: Kevin Sharp has been an engineer since long before he got his engineering degree, and has extensive experience in data acquisition and control networks in industrial, retail, and supply chain environments. He’s currently a freelance writer based in Tucson, Arizona.