12 ms·

How New Long-Range Radios Will Change the Internet of Things

186 points by slewis 11y ago

1024core 11y ago
> So now you can buy a radio chip for a few bucks and add it to any device. Where? > Think kilobits-per-second, not megabits-per-second. Well, voice is (or was...) 32kbps, so technically you could make a voice call over these radios, if they can give you 32kbps? I've heard of voice compression as low as 8kbps too. Imagine a cellphone that lasts months :)
- dconrad 11y ago
  http://www.digikey.com/catalog/en/partgroup/sx1272-and-sx1273/42224 http://www.digikey.com/catalog/en/partgroup/sx1272-and-sx127...
- jasonjayr 11y ago
  Codec2 is an open source speech codec, shooting for 700 - 3200 bit/s, specifically for digital voice over radio. http://www.rowetel.com/blog/?page_id=452 http://www.rowetel.com/blog/?page_id=452
- fsiefken 11y ago
  Opus SILK can go as low as 6000 bs. You can use it in existing VoIP solutions, VoIP over LoRa radio's would be very cool.. but I'm not sure if you can layer tcp/ip on top them yet.
th0ma5 11y ago
Lower power, but highly processed radio signals are fascinating to me. I've been playing with WSPR on the 10 meter band which is fixing (hopefully) to open up a little here over the next couple of months. With about 1/10th of the power of my Wi-Fi router and a long wire over top of the house, I've been picked up so far this year in California from Ohio. Last year I was heard almost daily in Europe. A blogger I follow with a bit more power but still less than a watt has been heard on several continents almost every other day from the UK. Anyway, signals are a function of time as well as bandwidth, and with more precise clocks in our devices and higher CPU power to make sense of it, it is an exciting time live and see what's next.
- anderiv 11y ago
  One of my most memorable moments in amateur radio was making a JT65 contact - 1 watt, from the north-central US to the island of Reunion, roughly 10k Kilometers. And that was on a very ad-hoc wire antenna.
- olympus 11y ago
  What's your callsign? I'll see if I can pick you up with my RTL-SDR+Ham-it-up and maybe once I have my ham license I'll talk back. Every now and then I find a post on HN that really makes my day and I realize how awesome this community is and how our interests overlap beyond computer programming and entrepreneurship. This is one of them.
  - kc2zuf 11y ago
    Not the OP but I'm on WSPR on 20m right now as KC2ZUF!
  - th0ma5 11y ago
    Well, I'd appreciate a report on the WSPRNet page once you get your ticket. I'm W8KV, but I don't know that I'm putting out all that much, and when I look at streetview of the houses of the people that pick up the signal they have some crazy setups! That said, I was able to receive PSK31 from France here in Ohio back before I got my ticket with just an old AC extension cord strung along the ceiling as an antenna, so keep playing with things, you never know what all will work.
- semi-extrinsic 11y ago
  Ok, since there are some radio wizards in this thread: if I want to log low bandwidth sensor data (so < 1kbps) on a unidirectional (if that helps) link over ~ 500m, almost clean line-of-sight with some trees in there, what are my options? I'd prefer to spend little money and use little power on the sensor end. I'm considering an ESP8266 (so 802.11) with a high-gain directional antenna. Sender and receiver are fixed locations.
  - olympus 11y ago
jws 11y ago
I'm intrigued. Anyone want to suggest an economical development board or transceiver?
- olympus 11y ago
  Like the author mentioned, these are new-ish/emerging standards and there isn't a de-facto board like Arduino just yet. Therefore it's kind of hard to recommend a board that I know has a ton of documentation and a big community. However, if you are the adventurous type and don't mind working with unpolished libraries and figuring things out from function descriptions, there are tons of smaller people putting out boards on Tindie: https://www.tindie.com/ https://www.tindie.com/ Tindie has a few LoRa boards, some with built in antennas and some without: https://www.tindie.com/products/DORJI_COM/long-range-semtech-lora-sx1276-sx1278-data-radio-modem-drf1278dm/?pt=full_prod_search https://www.tindie.com/products/DORJI_COM/long-range-semtech... But bottom line is that this tech isn't (yet) popular enough to try without getting a little dirty. If you can afford a bigger battery, try 900 MHz Xbee transmitters for an easy to interface serial with long range.
- dconrad 11y ago
  For LoRa, check out Multitech, LinkLabs, and Libelium. Semtech has a development board, too, though it would need to be paired with a microcontroller. Warning that most of this stuff is not for the faint of heart -- you need to be ready to dive into datasheets, bring up radio drivers, etc. For Sigfox, if you're not in an area with a Sigfox network, there's not much you can do. But TI has some killer FSK dev boards. http://www.ti.com/tool/smartrftrxebk http://www.ti.com/tool/smartrftrxebk and http://www.ti.com/tool/cc1101cc1190emk915 http://www.ti.com/tool/cc1101cc1190emk915 Ingenu has radio modules and dev boards, but I don't think they're set up to sell single-unit quantity. If you're in San Francisco, we're making a going to be distributing dev boards that will work out of the box and come with a bunch of sensors. Just sign up on our website: www.beepnetworks.com
  - kazazes 11y ago
    If you haven't lost your faith in Kickstarter yet, this project has $180k committed. A €40 pledge will get you a LoRa board and €200 for a LoRa gateway and what is essentially a demo dongle. https://www.kickstarter.com/projects/419277966/the-things-network https://www.kickstarter.com/projects/419277966/the-things-ne... Edit: fat thumbs, on mobile.
JupiterMoon 11y ago
Would this open possibilities for a direct machine to machine network that didn't go over the internet? (I.e. could I set up something to ssh to my home server from my laptop without needing wifi/cellular coverage?
- upofadown 11y ago
  The data rate is really low. So maybe, if you are patient and are not much more than a mile from home.
  - kevinnk 11y ago
    Data rate in the US is something like 11kbps. So you'd have to be really patient.
    - runeks 11y ago
      Isn't 11 Kbps fine for SSH? That's 11 thousand ascii characters per second. The latency is the important part, I assume this is low as well but I don't know.
      - olympus 11y ago
        Actually each ascii character takes up a byte (8 bits). You can get this down to just 7 bits if you aren't using the extended table, but a byte is very convenient and what most programs use for the size of a char. That means that with an 11Kbps line you are getting about 1,000 characters per second once you factor in other communication over head (stop bits, parity bits, etc). While this may sound fast it is painfully slow for anything other than typing in single line commands and reading single line responses. The transfer of a blank 4KB text file will take three seconds. The transfer of a plain text file the size of a book would take a few minutes. With the additional overhead of SSH you will actually see the characters appearing on the screen like a lame 80s hacker movie.
        marssaxman 11y ago
beardicus 11y ago
I just saw this kickstarter this morning, with related technology: https://www.kickstarter.com/projects/419277966/the-things-network?ref=balls https://www.kickstarter.com/projects/419277966/the-things-ne... They seem to have set up an open LoRa network in Amsterdam.
sbierwagen 11y ago
Yeah, but how big are the antennas?
- dconrad 11y ago
  Standard 800/900/2400 antennas on the client devices, obviously you can get more range out of a long high-gain antenna, but PCB antennas work as well.
blhack 11y ago
This is un. fucking. real. I work in the IoT space (specifically on health devices), and something like this would be absolutely a GOD SEND for the things that we're doing. The current "hottness" in our space is bluetooth low energy (BLE). This is super low power (or, rather: chips that are really good at going to deep sleep, and then coming online very quickly to burst some data out), but the range on BLE isn't great. Some companies use general purpose radios (these guys: http://www.glowcaps.com/ http://www.glowcaps.com/ are basically an couple of atmega 328s and some 433mhz radios) for range, but it seems like everything is moving towards BLE because you can use your cellphone/tablet as a base station. If LORA can seriously get this sort of range (I'm skeptical, honestly), then that is a total game-changer for IoT. A friend of mine at a chip manufacturer in town was telling me about some radio thing they were doing that had range like this about a year ago, but I didn't believe him. That little of power, for that long of range seems like straight-up magic. Super cool stuff! Very excited to see this really happening! Not to get too silly, but this about sums up how this makes me feel with regards to IoT: https://media.giphy.com/media/rl0FOxdz7CcxO/giphy.gif https://media.giphy.com/media/rl0FOxdz7CcxO/giphy.gif
- jazzyk 11y ago
  Cool stuff, sure. But, in my opinion, IoT will be slow to take off until the security aspect is solved - and it is a tough one.
  - dconrad 11y ago
    Security in the LoRaWAN standard is well thought-through, you might be surprised.
    - bdamm 11y ago
      Googling this did not lead to joy. Do you have a direct link?
      - slewis 11y ago
        https://www.lora-alliance.org/portals/0/specs/LoRaWAN%20Specification%201R0.pdf https://www.lora-alliance.org/portals/0/specs/LoRaWAN%20Spec... Would love to hear thoughts on it. Email address in my profile.
Joeri 11y ago
Our offices are equipped with sensors with LoRa radio's. We're partnering with a local cell service provider (proximus.be) who are upgrading their base stations to be LoRa base stations and we're a pilot installation because we're developing software that does things with the sensor data (smart office type stuff, like adaptive cleaning schedules based on space utilization). The sensors are pre-configured for their network. You stick them to the wall, pull a latch, and they connect directly to the public base station, no additional hardware involved. Surprisingly, it works, and they're claiming five years battery life. There are some downsides though. The sensor+radio combo's are custom orders for now, meaning they take too long to arrive and are too expensive. That issue will get solved over time. They also have an extremely low data rate. A single set of values every 15 minutes. That issue probably won't be solved. Still, you can learn a lot from a CO2 reading or a door opening counter reading even if it's only every 15 minutes. The hard bit is the intelligent software that processes the data. If anything is the achilles heel of IoT it's that lots of data does not necessarily lead to any insights.
bdamm 11y ago
Don't get too excited. Low-power devices get really interesting when they start to scale up in numbers. None of the technologies listed in this article have demonstrated that capability. The coming IoT shakedown is going to leave quite a few ebullient "visionaries" crying in their martini while they consider their next startup. * Star networks are infrastructure heavy, requiring build-outs on part with cellular networks, and having low per-base density. * LoRa's CSS (Chirp Spread Spectrum) approach doesn't promise to scale well, as devices start to interfere with each other over its very wide spectrum scattershot. There are other problems. Disclaimer: I work for an IoT mesh company (SSNI/WiSun) and we're busy cranking out massive networks that actually work. Sadly you can't get a dev kit at Fry's yet.
- slewis 11y ago
  Shawn from BeepNetworks here. Check out OnRamp (now Ingenu). Their coding scheme allows 1600 simultaneously communicating devices per spread factor per channel. That 1600 factor means their base station can handle massive numbers of clients compared to cellular. http://www.scribd.com/doc/273814075/On-Ramp-Wireless-White-Paper#scribd http://www.scribd.com/doc/273814075/On-Ramp-Wireless-White-P.... That paper is certainly biased in their favor, but the tech is real and they have serious deployments. The counterpoint to mesh is that it's a) much more complex to implement b) uses more power since nodes need to relay each others' communications. It will be interesting to see how it plays out, and who ends up crying in whose martinis as you say.
  - wyldfire 11y ago
    Hm, 1600 sounds like a lot. But depending on the applications it sounds like it might not be that much if it's 1600 over three square miles.
    - dconrad 11y ago
      that's 1600 devices simultaneously communicating. a sensor might only spend a fraction of a percent of its time transmitting data, so the number of devices supported is orders of magnitude higher.
bertm 11y ago
LoRa is a really cool technology. We are incorporating it into our networking stack/ hardware module. From field testing, we've seen about 4-6 miles line of sight, and about a mile through an urban environment (parking garage, apartment building, etc). Plus the bare radio chips are less than $10 bucks.
icholy 11y ago
I wonder how it will scale. Sure, you can use spread spectrum but to other unlicensed users it will look like the noise floor is increasing. As more people connect, the more noise there will be, and the players will either increase their power to maintain their SNRs, or have to slow down their rates.
TeMPOraL 11y ago
> So now you can buy a radio chip for a few bucks and add it to any device. Can someone tell me some names of those radio chips? I'm in Shenzhen, if they exist they should be available here. I want to buy some for a few bucks.
- dmritard96 11y ago
  Agreed. Seriously doubt there is some fundamentally new chips or anything. My suspicion is like, a cc430 + a front end radio amp lol
- jasonjs 11y ago
  These radios have been available for quite some time, it's the Semtech SX127x line (e.g. SX1276). I was looking into LoRa and other long-range IoT WAN radios about a year ago for a research project. You can get cheap (22 USD) Arduino-compatible development boards here using HopeRF modules, which work great with the RadioHead library: http://www.anarduino.com/miniwireless/ http://www.anarduino.com/miniwireless/ http://www.airspayce.com/mikem/arduino/RadioHead/ http://www.airspayce.com/mikem/arduino/RadioHead/ (I have no affiliation with the websites.)
- nicolsc 11y ago
  Sigfox-ready chips can be bought from Atmel, Axsem, SiLabs, TI, .. and more : http://makers.sigfox.com/ http://makers.sigfox.com/ You should be able to get some Sigfox-enabled TI CC1120 for a few $ easily.
Animats 11y ago
The main innovation here is battery life. You can do most of these things now with two-way pager technology.[1] There are a lot of commercial air conditioning units and vending machines phoning home that way right now. They all have AC power, so battery life isn't an issue. Two way pagers have more or less died out for humans, but for M2M (machine to machine) communication they are alive and well. [1] https://en.wikipedia.org/wiki/ReFLEX https://en.wikipedia.org/wiki/ReFLEX
KaiserPro 11y ago
if anyone wants to play with LoRa you might want to play with the RFM 69 http://rdepablos.merlitec.com/mixed/rfm69-library-for-raspberry-pi http://rdepablos.merlitec.com/mixed/rfm69-library-for-raspbe... Gets about 1.6 miles with basic wire antenna. More if you play with directional yagis or the like The problem with LoRa is that yes you can get ten miles, assuming you chose the right antenna and radio, but you wont be able to transmit for very long. I have a number of sensors that have RFM29 radios in them, and they run off solar and super caps. However they only transmit once a minute, and they don't stay on for a reply. Power management is key at low power levels. To get ten years battery life, you either need massive batteries, or only transmit once every day/hour.
- jasonjs 11y ago
  The RFM69 does not use LoRa, maybe you meant the RFM95W/96W?
  - KaiserPro 11y ago
    Sorry yes. The 69 is a packet radio, which gives you more freedom, in the same way raw ethernet does.... :)
andyl 11y ago
Looks like LoRa is star topology. Is there another comparable tech that supports mesh networking?
boothead 11y ago
Fun fact: spread spectrum was invented by a hollywood actress during WWII http://www.women-inventors.com/Hedy-Lammar.asp http://www.women-inventors.com/Hedy-Lammar.asp
s3nnyy 11y ago
There is a lorawan bootcamp soon coming up in San Jose (https://www.regonline.com/register/checkin.aspx?EventId=1732983&MethodId=0&EventSessionId=&startnewreg=1 https://www.regonline.com/register/checkin.aspx?EventId=1732...). I attended this bootcamp in Zurich, Switzerland a couple of days ago and it was quite interesting. Swisscom, the biggest Telco here covered Zurich and Geneva already with a lorawan net and they will give access to it during an upcoming hackathon (http://iot-hackathon.swisscom.com/ http://iot-hackathon.swisscom.com/).
fit2rule 11y ago
You know what will really change the IoT? Energy-harvesting. (As soon as the powers that be stop interfering with it at the foundry level, that is..)

186 points by slewis 11y ago

> So now you can buy a radio chip for a few bucks and add it to any device. Where? > Think kilobits-per-second, not megabits-per-second. Well, voice is (or was...) 32kbps, so technically you could make a voice call over these radios, if they can give you 32kbps? I've heard of voice compression as low as 8kbps too. Imagine a cellphone that lasts months :)

dconrad 11y ago
http://www.digikey.com/catalog/en/partgroup/sx1272-and-sx1273/42224 http://www.digikey.com/catalog/en/partgroup/sx1272-and-sx127...
jasonjayr 11y ago
Codec2 is an open source speech codec, shooting for 700 - 3200 bit/s, specifically for digital voice over radio. http://www.rowetel.com/blog/?page_id=452 http://www.rowetel.com/blog/?page_id=452
fsiefken 11y ago
Opus SILK can go as low as 6000 bs. You can use it in existing VoIP solutions, VoIP over LoRa radio's would be very cool.. but I'm not sure if you can layer tcp/ip on top them yet.

th0ma5 11y ago

Lower power, but highly processed radio signals are fascinating to me. I've been playing with WSPR on the 10 meter band which is fixing (hopefully) to open up a little here over the next couple of months. With about 1/10th of the power of my Wi-Fi router and a long wire over top of the house, I've been picked up so far this year in California from Ohio. Last year I was heard almost daily in Europe. A blogger I follow with a bit more power but still less than a watt has been heard on several continents almost every other day from the UK. Anyway, signals are a function of time as well as bandwidth, and with more precise clocks in our devices and higher CPU power to make sense of it, it is an exciting time live and see what's next.

anderiv 11y ago
One of my most memorable moments in amateur radio was making a JT65 contact - 1 watt, from the north-central US to the island of Reunion, roughly 10k Kilometers. And that was on a very ad-hoc wire antenna.
olympus 11y ago
What's your callsign? I'll see if I can pick you up with my RTL-SDR+Ham-it-up and maybe once I have my ham license I'll talk back. Every now and then I find a post on HN that really makes my day and I realize how awesome this community is and how our interests overlap beyond computer programming and entrepreneurship. This is one of them.
- kc2zuf 11y ago
  Not the OP but I'm on WSPR on 20m right now as KC2ZUF!
- th0ma5 11y ago
  Well, I'd appreciate a report on the WSPRNet page once you get your ticket. I'm W8KV, but I don't know that I'm putting out all that much, and when I look at streetview of the houses of the people that pick up the signal they have some crazy setups! That said, I was able to receive PSK31 from France here in Ohio back before I got my ticket with just an old AC extension cord strung along the ceiling as an antenna, so keep playing with things, you never know what all will work.
semi-extrinsic 11y ago
Ok, since there are some radio wizards in this thread: if I want to log low bandwidth sensor data (so < 1kbps) on a unidirectional (if that helps) link over ~ 500m, almost clean line-of-sight with some trees in there, what are my options? I'd prefer to spend little money and use little power on the sensor end. I'm considering an ESP8266 (so 802.11) with a high-gain directional antenna. Sender and receiver are fixed locations.
- olympus 11y ago

jws 11y ago

I'm intrigued. Anyone want to suggest an economical development board or transceiver?

olympus 11y ago
Like the author mentioned, these are new-ish/emerging standards and there isn't a de-facto board like Arduino just yet. Therefore it's kind of hard to recommend a board that I know has a ton of documentation and a big community. However, if you are the adventurous type and don't mind working with unpolished libraries and figuring things out from function descriptions, there are tons of smaller people putting out boards on Tindie: https://www.tindie.com/ https://www.tindie.com/ Tindie has a few LoRa boards, some with built in antennas and some without: https://www.tindie.com/products/DORJI_COM/long-range-semtech-lora-sx1276-sx1278-data-radio-modem-drf1278dm/?pt=full_prod_search https://www.tindie.com/products/DORJI_COM/long-range-semtech... But bottom line is that this tech isn't (yet) popular enough to try without getting a little dirty. If you can afford a bigger battery, try 900 MHz Xbee transmitters for an easy to interface serial with long range.
dconrad 11y ago
For LoRa, check out Multitech, LinkLabs, and Libelium. Semtech has a development board, too, though it would need to be paired with a microcontroller. Warning that most of this stuff is not for the faint of heart -- you need to be ready to dive into datasheets, bring up radio drivers, etc. For Sigfox, if you're not in an area with a Sigfox network, there's not much you can do. But TI has some killer FSK dev boards. http://www.ti.com/tool/smartrftrxebk http://www.ti.com/tool/smartrftrxebk and http://www.ti.com/tool/cc1101cc1190emk915 http://www.ti.com/tool/cc1101cc1190emk915 Ingenu has radio modules and dev boards, but I don't think they're set up to sell single-unit quantity. If you're in San Francisco, we're making a going to be distributing dev boards that will work out of the box and come with a bunch of sensors. Just sign up on our website: www.beepnetworks.com
- kazazes 11y ago
  If you haven't lost your faith in Kickstarter yet, this project has $180k committed. A €40 pledge will get you a LoRa board and €200 for a LoRa gateway and what is essentially a demo dongle. https://www.kickstarter.com/projects/419277966/the-things-network https://www.kickstarter.com/projects/419277966/the-things-ne... Edit: fat thumbs, on mobile.

JupiterMoon 11y ago

Would this open possibilities for a direct machine to machine network that didn't go over the internet? (I.e. could I set up something to ssh to my home server from my laptop without needing wifi/cellular coverage?

upofadown 11y ago
The data rate is really low. So maybe, if you are patient and are not much more than a mile from home.
- kevinnk 11y ago
  Data rate in the US is something like 11kbps. So you'd have to be really patient.
  - runeks 11y ago
    Isn't 11 Kbps fine for SSH? That's 11 thousand ascii characters per second. The latency is the important part, I assume this is low as well but I don't know.
    - olympus 11y ago
      Actually each ascii character takes up a byte (8 bits). You can get this down to just 7 bits if you aren't using the extended table, but a byte is very convenient and what most programs use for the size of a char. That means that with an 11Kbps line you are getting about 1,000 characters per second once you factor in other communication over head (stop bits, parity bits, etc). While this may sound fast it is painfully slow for anything other than typing in single line commands and reading single line responses. The transfer of a blank 4KB text file will take three seconds. The transfer of a plain text file the size of a book would take a few minutes. With the additional overhead of SSH you will actually see the characters appearing on the screen like a lame 80s hacker movie.
      - marssaxman 11y ago

beardicus 11y ago

I just saw this kickstarter this morning, with related technology: https://www.kickstarter.com/projects/419277966/the-things-network?ref=balls https://www.kickstarter.com/projects/419277966/the-things-ne... They seem to have set up an open LoRa network in Amsterdam.

sbierwagen 11y ago

Yeah, but how big are the antennas?

dconrad 11y ago
Standard 800/900/2400 antennas on the client devices, obviously you can get more range out of a long high-gain antenna, but PCB antennas work as well.

blhack 11y ago

This is un. fucking. real. I work in the IoT space (specifically on health devices), and something like this would be absolutely a GOD SEND for the things that we're doing. The current "hottness" in our space is bluetooth low energy (BLE). This is super low power (or, rather: chips that are really good at going to deep sleep, and then coming online very quickly to burst some data out), but the range on BLE isn't great. Some companies use general purpose radios (these guys: http://www.glowcaps.com/ http://www.glowcaps.com/ are basically an couple of atmega 328s and some 433mhz radios) for range, but it seems like everything is moving towards BLE because you can use your cellphone/tablet as a base station. If LORA can seriously get this sort of range (I'm skeptical, honestly), then that is a total game-changer for IoT. A friend of mine at a chip manufacturer in town was telling me about some radio thing they were doing that had range like this about a year ago, but I didn't believe him. That little of power, for that long of range seems like straight-up magic. Super cool stuff! Very excited to see this really happening! Not to get too silly, but this about sums up how this makes me feel with regards to IoT: https://media.giphy.com/media/rl0FOxdz7CcxO/giphy.gif https://media.giphy.com/media/rl0FOxdz7CcxO/giphy.gif

jazzyk 11y ago
Cool stuff, sure. But, in my opinion, IoT will be slow to take off until the security aspect is solved - and it is a tough one.
- dconrad 11y ago
  Security in the LoRaWAN standard is well thought-through, you might be surprised.
  - bdamm 11y ago
    Googling this did not lead to joy. Do you have a direct link?
    - slewis 11y ago
      https://www.lora-alliance.org/portals/0/specs/LoRaWAN%20Specification%201R0.pdf https://www.lora-alliance.org/portals/0/specs/LoRaWAN%20Spec... Would love to hear thoughts on it. Email address in my profile.

Joeri 11y ago

Our offices are equipped with sensors with LoRa radio's. We're partnering with a local cell service provider (proximus.be) who are upgrading their base stations to be LoRa base stations and we're a pilot installation because we're developing software that does things with the sensor data (smart office type stuff, like adaptive cleaning schedules based on space utilization). The sensors are pre-configured for their network. You stick them to the wall, pull a latch, and they connect directly to the public base station, no additional hardware involved. Surprisingly, it works, and they're claiming five years battery life. There are some downsides though. The sensor+radio combo's are custom orders for now, meaning they take too long to arrive and are too expensive. That issue will get solved over time. They also have an extremely low data rate. A single set of values every 15 minutes. That issue probably won't be solved. Still, you can learn a lot from a CO2 reading or a door opening counter reading even if it's only every 15 minutes. The hard bit is the intelligent software that processes the data. If anything is the achilles heel of IoT it's that lots of data does not necessarily lead to any insights.

bdamm 11y ago

Don't get too excited. Low-power devices get really interesting when they start to scale up in numbers. None of the technologies listed in this article have demonstrated that capability. The coming IoT shakedown is going to leave quite a few ebullient "visionaries" crying in their martini while they consider their next startup. * Star networks are infrastructure heavy, requiring build-outs on part with cellular networks, and having low per-base density. * LoRa's CSS (Chirp Spread Spectrum) approach doesn't promise to scale well, as devices start to interfere with each other over its very wide spectrum scattershot. There are other problems. Disclaimer: I work for an IoT mesh company (SSNI/WiSun) and we're busy cranking out massive networks that actually work. Sadly you can't get a dev kit at Fry's yet.

slewis 11y ago
Shawn from BeepNetworks here. Check out OnRamp (now Ingenu). Their coding scheme allows 1600 simultaneously communicating devices per spread factor per channel. That 1600 factor means their base station can handle massive numbers of clients compared to cellular. http://www.scribd.com/doc/273814075/On-Ramp-Wireless-White-Paper#scribd http://www.scribd.com/doc/273814075/On-Ramp-Wireless-White-P.... That paper is certainly biased in their favor, but the tech is real and they have serious deployments. The counterpoint to mesh is that it's a) much more complex to implement b) uses more power since nodes need to relay each others' communications. It will be interesting to see how it plays out, and who ends up crying in whose martinis as you say.
- wyldfire 11y ago
  Hm, 1600 sounds like a lot. But depending on the applications it sounds like it might not be that much if it's 1600 over three square miles.
  - dconrad 11y ago
    that's 1600 devices simultaneously communicating. a sensor might only spend a fraction of a percent of its time transmitting data, so the number of devices supported is orders of magnitude higher.

bertm 11y ago

LoRa is a really cool technology. We are incorporating it into our networking stack/ hardware module. From field testing, we've seen about 4-6 miles line of sight, and about a mile through an urban environment (parking garage, apartment building, etc). Plus the bare radio chips are less than $10 bucks.

icholy 11y ago

I wonder how it will scale. Sure, you can use spread spectrum but to other unlicensed users it will look like the noise floor is increasing. As more people connect, the more noise there will be, and the players will either increase their power to maintain their SNRs, or have to slow down their rates.

TeMPOraL 11y ago

> So now you can buy a radio chip for a few bucks and add it to any device. Can someone tell me some names of those radio chips? I'm in Shenzhen, if they exist they should be available here. I want to buy some for a few bucks.

dmritard96 11y ago
Agreed. Seriously doubt there is some fundamentally new chips or anything. My suspicion is like, a cc430 + a front end radio amp lol
jasonjs 11y ago
These radios have been available for quite some time, it's the Semtech SX127x line (e.g. SX1276). I was looking into LoRa and other long-range IoT WAN radios about a year ago for a research project. You can get cheap (22 USD) Arduino-compatible development boards here using HopeRF modules, which work great with the RadioHead library: http://www.anarduino.com/miniwireless/ http://www.anarduino.com/miniwireless/ http://www.airspayce.com/mikem/arduino/RadioHead/ http://www.airspayce.com/mikem/arduino/RadioHead/ (I have no affiliation with the websites.)
nicolsc 11y ago
Sigfox-ready chips can be bought from Atmel, Axsem, SiLabs, TI, .. and more : http://makers.sigfox.com/ http://makers.sigfox.com/ You should be able to get some Sigfox-enabled TI CC1120 for a few $ easily.

Animats 11y ago

The main innovation here is battery life. You can do most of these things now with two-way pager technology.[1] There are a lot of commercial air conditioning units and vending machines phoning home that way right now. They all have AC power, so battery life isn't an issue. Two way pagers have more or less died out for humans, but for M2M (machine to machine) communication they are alive and well. [1] https://en.wikipedia.org/wiki/ReFLEX https://en.wikipedia.org/wiki/ReFLEX

KaiserPro 11y ago

if anyone wants to play with LoRa you might want to play with the RFM 69 http://rdepablos.merlitec.com/mixed/rfm69-library-for-raspberry-pi http://rdepablos.merlitec.com/mixed/rfm69-library-for-raspbe... Gets about 1.6 miles with basic wire antenna. More if you play with directional yagis or the like The problem with LoRa is that yes you can get ten miles, assuming you chose the right antenna and radio, but you wont be able to transmit for very long. I have a number of sensors that have RFM29 radios in them, and they run off solar and super caps. However they only transmit once a minute, and they don't stay on for a reply. Power management is key at low power levels. To get ten years battery life, you either need massive batteries, or only transmit once every day/hour.

jasonjs 11y ago
The RFM69 does not use LoRa, maybe you meant the RFM95W/96W?
- KaiserPro 11y ago
  Sorry yes. The 69 is a packet radio, which gives you more freedom, in the same way raw ethernet does.... :)

andyl 11y ago

Looks like LoRa is star topology. Is there another comparable tech that supports mesh networking?

boothead 11y ago

Fun fact: spread spectrum was invented by a hollywood actress during WWII http://www.women-inventors.com/Hedy-Lammar.asp http://www.women-inventors.com/Hedy-Lammar.asp

s3nnyy 11y ago

There is a lorawan bootcamp soon coming up in San Jose (https://www.regonline.com/register/checkin.aspx?EventId=1732983&MethodId=0&EventSessionId=&startnewreg=1 https://www.regonline.com/register/checkin.aspx?EventId=1732...). I attended this bootcamp in Zurich, Switzerland a couple of days ago and it was quite interesting. Swisscom, the biggest Telco here covered Zurich and Geneva already with a lorawan net and they will give access to it during an upcoming hackathon (http://iot-hackathon.swisscom.com/ http://iot-hackathon.swisscom.com/).

fit2rule 11y ago

You know what will really change the IoT? Energy-harvesting. (As soon as the powers that be stop interfering with it at the foundry level, that is..)