Table Comparing Wireless Protocols for IoT Devices

We compiled this table to help our team select the best wireless communication protocol for the products we make. We hope it helps you as much as it has helped us.

Wireless Protocols for IoTFrequencyRange (non-urban environment)Data RatePower DrawTopologyRequires Hub or Gateway?Proprietary or Open?Chipset CostManaged BySecurityIntended UseOur Thoughts
ZigBee2.4GHz, 915MHz (US), 868 MHz (EU)100-325 ft250 kbps (2.4) 40kbps (915) 20kbps (868)lowMeshYesOpen$$ZigBee Alliance (Comcast, Kroger, Samsung, TI)EncryptedSingle BuildingZigbee is fractured, has multiple standards and one Zigbee product may not work with another Zigbee product. Zigbee can only be managed by one controller. Despite these shortcomings, ZigBee is already installed across the globe and ZigBee chips are readily available to develop on. For home-targeted products that need to be rolled out quickly, ZigBee is a good choice. ZigBee has the same downside as other mesh networks: many devices are required for reliable operation, and latency is relatively high. The upside is that ZigBee is already widely available and adopted, and works well if all of your ZigBee devices use the same standard. We think it's a good choice for low-cost products targeted at the home.
Z-Wave915MHz (US) 868MHz (EU)100-325 ft40kbps (915) 20kbps (868)lowMeshYesProprietary$$Z-Wave AllianceEncryptedSingle BuildingZ-Wave is less fractured than ZigBee and probably has a better market share of Home Depot/Best Buy products in the U.S. Z-Wave uses a lower frequency than US ZigBee, which means it should have better range and less power draw. Unfortunately, Z-Wave chips are generally more expensive because they are only made by one manufacturer. Z-Wave is mesh, so many devices are required for reliable operation and latency is relatively high. Despite these shortcomings, Z-Wave is widely adopted and reliable if setup correctly. We think it's a good choice for products targeted at the home.
Bluetooth 4.0+2.4GHz200 ft25MbpsmediumPANYesOpen$Bluetooth Special Interest Group (3k members)EncryptedPersonalBluetooth is tempting to use for IoT products because it is built into every smartphone already. Bluetooth can have high data rates and low power consumption. The downside of Bluetooth for IoT is the PAN network model. It's challenging to have multiple devices on the network. There is generally a limit of 8 devices. Bluetooth 4.0+ is is a good choice for products that can be managed from just a smartphone.
Bluetooth 52.4GHz800 ft50MbpsmediumPANYesOpen$Bluetooth Special Interest Group (3k members)EncryptedPersonalThe latest Bluetooth standard should have 4x the range and 2x the data rate of Bluetooth 4.0+. It still uses the PAN network model so it shares many of the same challenges as Bluetooth 4.0+. If you're going to start building a Bluetooth product, you should include the latest standard Bluetooth chip which will be shipping in all smartphones soon.
Bluetooth Low Energy (BLE)2.4GHz200 ft10kB/slowPANYesOpen$Bluetooth Special Interest Group (3k members)EncryptedPersonalBLE is essentially Bluetooth except it goes into sleep mode after connecting for a few mS. The low power consumption means the BLE is a better protocol for IoT, except it still uses the very limited PAN network topology. There is an upcoming BLE Mesh standard which SHOULD fix the issues of the PAN network model. If it does, BLE will be a very powerful IoT communication protocol. BLE is already a great technology choice for wearable products.
Wi-Fi2.4GHZ/5GHz115-2307GbpshighStarNoOpen$$IEEEOptionalSingle BuildingWi-Fi is readily available in most commercial buildings and homes. This is a massive advantage for IoT products targeting those markets. Because of the pre-existing Wi-Fi networks, Wi-Fi products do not require a hub that's separate from the router. They don't need unreliable mesh networks to extend range either. They have instant access to the cloud. The downside of Wi-Fi is that it can be difficult for the consumer to get it connected to their router and it has a very high power draw. Wi-Fi is a great technology choice for standalone products targeted at the home or business. It can be used for battery-powered products if power is managed appropriately.
Wi-Fi-ah (HaLow)900MHz3000 ft347MbpslowStarNoOpen???IEEE???Single BuildingHaLow requires a special Wi-Fi router that's available on the market now but not installed in most homes. HaLow devices will have instant internet access like traditional Wi-Fi devices assuming the router is HaLow compatible. Halow has better wall penetration and range than Wi-Fi because it uses the lower 900MHz frequency band. This also means lower power draw for battery operated devices. If all routers start shipping with HaLow built in, this will be a very strong wireless protocol for homes and commercial buildings. Watch the adoption rate of HaLow, and plan your development accordingly.
Thread2.4GHz100 ft250kbpslowMeshYesOpen$$Thread Group (Google, Samsung, etc.)EncryptedSingle BuildingThread is backed by Google, so you know there are some excellent engineers behind it. It's built on the 6LoWPAN stack which uses the same 802.15.4 radio as ZigBee and could become a dominant player in the Home Automation space. It's seems to be built so that Nest becomes the ZigBee-like hub of the house. Nest uses it's Wi-Fi connection to get low-power thread devices online. It's a great idea in theory, but has yet to become widely adopted. There's discussion of building Thread compatibility with ZigBee devices. Thread is a great technology for home-targeted products that target customers who already have a Nest.
DigiMesh2.4GHz/900 MHz (US)/868 MHz (EU)~20 miles250 kbps (2.4) 40kbps (915) 20kbps (868)lowMeshYesProprietary$$$DigiMeshEncryptedSingle Building or WANDigiMesh is essentially modified ZigBee focused on long range point to point communication. It looks like a good technology, but nobody that I found has actually built products on it. This may be because the cost per chip is very high. For products targeted at large commercial buildings without per-device cost constraints, DigiMesh is a good choice.
MiWi2.4GHz or subGHz800 ft250kbpslowMesh or StarYesProprietary$MiWiEncryptedSingle Building or WANMiWi is similar to DigiMesh in that it's a modified and proprietary form of ZigBee. It requires less power, lower memory, and is good for very low-cost products and systems. It's not widely adopted yet but could be good for products that will require a custom hub anyway. MiWi also has low memory constraints, which makes it a good choice for products that have to have a very low per-unit BOM cost.
EnOcean900Mhz (US) 868 MHz (EU) 315 MHz30-100 ft125kbps"Battery Free"MeshYesProprietary$$EnOceanEncryptedSingle BuildingBattery-free operation promises a long device life. It's a very interesting technology that we haven't had a chance to play with, but we follow the company closely. EnOcean can be prototyped with a raspberry pi which lowers development costs. EnOcean is a good technology choice for products targeted at the commercial building that should have low maintenance costs.
6LoWPAN2.4GHz380 ft250kbpslowMeshYesOpen$IEEEOptionalSingle Building6LoWPAN is a promising alternative to other mesh network technologies. Because it's based on IPV6 addressing, it's relatively simple for 6LoWPAN devices to communicate with other IoT networks by building a bridge. For example, a 6LoWPAN to Wi-Fi bridge is simpler to produce and operate than a Zigbee hub. In theory, the 6LoWPAN devices would have almost direct access to the Wi-Fi devices. 6LoWPAN is another standard that's great in specific applications. We recommend it for products targeted at the home or commercial buildings that need to communicate with other products or systems.
Weightless (W, N, P)white-spaces, 915MHz, 868MHz, 780MHz, 470 MHz, 433 MHz, 169 MHz1.2 miles (P), 3 miles (W, N)200bps-100kbpslow (N), medium (W, P)StarYesOpen$$Weightless Special Interest GroupEncryptedWANWeightles-W was rejected by the FCC and other governing bodies, but N and P look like solid WAN technologies. Weightless is a SIG with tons of members and competeing ine the LPWAN space. Weightless N is one-way communication which is very limiting. Weightless-P looks like a great LPWAN technology, but it hasn't been deployed yet. With royalty free deployment, Weightless-P and Weightless-N look like good technologies for LPWAN products.
mcThings2.4GHz650 ft50kbpslowStarYesProprietary$$$mcThings???Single BuildingmcThings is great for deploying a custom set of sensors in a few buildings. The cost per unit is high, but the technology is very power efficient and requires little maintenance. You can easily expand a mcThings network with bridges, and battery life for basic sensors can be up to 10 years. We recommend mcThings for sensor-based products targeting a few businesses buildings.
LoRa150MHz-1GHz (lots of options)up to 20 miles50kbpslowStarKind ofOpen$$$LoRa AllianceEncryptedWANLoRaWAN is an alliance focused on creating a LPWAN technology for IoT devices. LoRa uses spread-spectrum technology that lets the LoRa chip decide the best spectrum to use for data rates, interference, and battery life. It's strongly adopted and deployed, with multiple vendors selling proven LoRa hardware. Because it's relatively inexpensive to cover a new area with LoRa, it's a good technology choice for LPWAN IoT products that need to be placed in areas without cell service.
SigFox900Mhz (US) 868 MHz (EU)~20 miles100bpslowStarYesProprietary$$$SigFoxEncryptedWANThe original player in the LPWAN space, SigFox had the vision to see LPWAN coming and has already deployed their network over most of Europe. SigFox is a proprietary technology, so your price per chip is relatively high. SigFox has great coverage right now, but they are threatened by the onset of LTE Cat M1 and NB-IoT. SigFox is a good choice for LPWAN products that need to be deployed in Europe right now.
LTE Cat-M11.4MHz~20 miles1MbpslowStarNoOpen???3GPP, LTE-M TaskForce???WANLTE M1 is not available yet, but it's a very exciting LPWAN technology. M1 should be deployable on existing LTE networks without hardware upgrades. That means the Verizon and AT&T could cover most of the US with LTE M1 with just a software upgrade (and both have announced plans to do just that). M1 has a high data rate, but devices are capable of sleeping to reduce power. We don't know what the power consumption will look like until we get our hands on a working M1 radio, but watch this technology closely. It's a major threat to LoRa and SigFox, which require the installation of new radio towers to deploy coverage. It could be a very good choice for products that target massive areas like nations, states, or cities.
NarrowBand-IoT (Cat M2)Below 1GHz~20 miles100kbpslowStarNoOpen???3GPP, Ericsson, HuaweiEncryptedWANNB-IoT is similar to LTE Cat M1, except it is GSM-based. NB-IoT till do better globally (where LTE networks are not prevalent) and well on T-Mobile & Sprint in the US. All these IoT technologies sleep but NB-IoT also usesless power than the competitors when the radio is on due to a relatively simple waveform. NB-IOT should also have a cheaper chip than it's LTE-M1 counterpart. Not officially rolled out yet, but it's being tested in a few areas and should be watched closely. It could be a good choice for products that target massive areas like nations, states, or cities.
3G and 4G Cellular (US)700 MHz, 800 MHz, 850MHz, 1700MHz, 1900MHz, 2100MHz, 2300MHz, 2500MHz~20 miles200kbps (3G) 10Mbps (4G)highStarNoOpen$$3GPPEncryptedWANCellular technology is not designed for IoT, but it's already rolled out across most of the globe. For IoT devices that do not require battery power and need to be launched immediately, cellular is a good choice. For IoT products that can wait to launch, it's worth waiting to see who comes out on top in the cellular LPWAN war.

Our Thoughts:

There are a ton of options for getting your IoT device connected wirelessly, and there is no clear winner for every case. We put together a few simple guidelines, please reach out to us for more detailed advice.

Frequency

Frequency plays a huge role in determining characteristics of a wireless protocol. Most IoT wireless protocols operate on unlicensed spectrum bands. This means that they don’t have to pay the FCC or similar governing body to broadcast on that band, but the risk of interference from other networks is high. In general, higher frequencies will have a higher data rate, lower range, higher power draw, and lower wall penetration. 2.4GHz is the only frequency that is unlicensed globally which is why it’s the de facto frequency for most IoT protocols.

Range

If your device is going to be installed in a home or building, looking for a lower frequency can help with wall penetration. If your device is going to be installed city-wide or state-wide, you want to find a device with WAN in the “Intended Use” column.

Power Draw

If your device is battery operated, you’re going to want to select a low-power protocol. Generally, higher frequencies draw more power. Some protocols can get around this with sleep modes (BLE) and altered MCUs to fit your use case, but frequency does play a role in power draw.

Data Requirements

As a frame of reference, most MP3 files stream at 256 kbps. Depending on your application, meeting the minimum data rate requirements is very important. Frequency once again plays a role here, higher frequency generally allows for a higher data rate.

Topology

Mesh networks are cheaper to install because they don’t require a powerful and centrally placed hub. But they do require a number of devices to function well, so they aren’t a good choice for standalone products. Star networks require some sort of reasonably powerful hub in a central location. But they make up for this cost with improved latency, data rates, and reliability (in some cases).