LoRa and Sigfox DISQUALIFIED...Why?
My inbox continues to pile up with the same questions over and over about the results of a massive, IoT tender to deploy over 200,000 end devices. So, I thought I would write a brief summary to share with my LinkedIn friends.
Not so often do we get to learn from a real, large-scale IoT deployment in which an objective party evaluates all wireless technologies suitable for a huge network of IoT devices. Here is what happened...
In a government conducted IoT tender spanning over 13 months, the Taiwan Power Company searched for the most cost effective, reliable wireless solutions to connect their end devices (smart meters).
The criteria was simple...
- Bi-Directional: End Devices can wirelessly communicate to the Base Station. Base Station can wirelessly communicate to the End Devices.
- Reliable: 99% success rate: end device sends usage data every 15 mins (20-40 bytes)
- Scalable: A single Base Station should provide multiple kilometers of coverage and connect thousands of end devices
LoRa's struggles: 1) unreliable downlink for key exchange, meter reconfiguration, message ACKS, firmware-over-the-air, etc 2) inter-network collisions, large usage of frequency bands = interference from others 3) low data rate = poor scalability for base stations to support end devices sending every 15 mins (20-40 bytes)
Sigfox's struggles: 1) pretty much obsolete downlink capability (max 4 messages a day) 2) max 0.6kbps makes it WAY too slow to handle requirements 3) overall, the furthest away from meeting any requirements
Ubiik used a new technology called Weightless to clinch victory; meeting all requirements and submitting a price advantageous bid. For this single project, Ubiik has already successfully installed over 70,000 end devices and over 300 base stations in a single network.
FAQ:
Q1:What is the point of smart meters sending every 15 min?
A1: It has become the norm of smart meter requisites. This allows electricity companies to make critical decisions based on energy usage big data (i.e whether to turn-on a power plant, how to quantify/adapt energy saving programs etc). In addition, energy theft costs reap billions of dollars annual losses (U.S alone estimates around 6 billion USD in losses to energy theft per year). With smart meters, each consumer will have historical, high resolution energy profiles that can be analyzed to make accurate assumptions on theft.
Q2: Who is Ubiik?
A2: An IoT solutions company. We provide full turn key solutions from conceptualization, industrial design, hardware manufacturing, and customized software solutions. We deliver end-to-end solutions with wireless technologies that deliver our partners the highest value and ROI: Weightless LPWAN, NB-IoT, BLE, and more...
Q3: How were tests conducted for the government tender?
Q3: Along with the mountains of paperwork, there were 3 main trials. #1: lab testing - 1,250 modules checked, 20 meters retrofitted sending every 15 min 99.8% reliability required. #2 Dense urban testing - 910 residential meters in Taipei City retrofitted with competing technologies for 7 days; 99% reliability required. #3 Taipower Shulin office: field trial , over 100 meters 99% reliability
Q4: What is Weightless?
Q4: Weightless is a narrowband low power, wide area network (LPWAN) technology achieving long range, low power, massive scalability and firmware-over-the-air (FOTA). Ubiik sells Weightless Base Stations and Weightless End Device modules.
More questions?...leave a comment and/or email me jay(at)ubiik.com
"You can expect no influence if you are not susceptible to influence." - Carl Jung
Founder at Wavelink Technology
5 年Congrats for all work and achievements ! and for nice marketing :) Reading the comments I See a lot of people making lot of confusions , due to lack of practical experience were the real needs meet solutions .? But also lots of well prepared professionals .... 1. The need of making uploads of load profiles on electricity side , so often(15 min like) is normal and? can be made by systems were? power consumption is not critical (always on grid), but were data volume can be big, or number of messages is? high (reliability of communications is important in this case). Usually number of messages is high when meter reports to head end system? by a scheduled communication. Its less needed a down link very often, or high volume data (upgrade FW over the air).? Such kind of requirements are hard to be met, by systems were same resource called spectrum, is used in a non-controlled manner? and we want to have Kilometers of coverage for one base station, and maybe its an urban area were one(1),? FREE, frequency band is abused by everybody ! CAT M1 could be easily the solution to such feature .(or another similar? and proprietary) 2. The non rechargeable battery devices (water and gas meters)? have different usage scenario were lower data are uploaded to the head end system (? maybe 100- 200 octets per day spread over 2-3 messages), BUT lifetime for the battery ,? for cases were? >15 years are required , is a big problem. Many consumer IoT companies are just simply calculating battery capacity? for a energetic profile. It is not that simple.? Battery over lifetime change ESR and therefore the voltage dips? during a transmission after long inactive time+passivization+ temperature histogram? are just simply changing the game.? You can never compare a system which takes never more than 50 mA in TX mode (even when antenna is not-tuned? Bad VSWR), with a system were TX takes in best conditions >=250 mA . (like NB Iot or CAt M1) Conclusion : for battery operated devices? were lifetime is important cannot be same technology as for ON Grid - high data rate, high data volumes , high reliability system. The trade off has to be made between the real need and technology.? Please stop to post comments to this nice article? , were some "guru" just points to best for everything technology , against some other technology? which is not good at all. All technologies have their role/market/use case scenario. Adapting a use case to the right technology, means less cost on the solution and less cost per part BoM. example: Doesn't make sense? to put an NB Iot module of at least 5$? were you need 15 years lifetime in all conditions and specific target price if you can do it wit LoRa or Sigfox at lower battery /more years of work. In the same time its hard to get reliable communication on this ones? and you cannot make high data-rate. This is the real diversity...
Research Cyber Security Team Lead
5 年Congrats Jay Wey Just a comment the issues are specific to LoRaWAN, LoRa as RF link has no such limitations. LoRa with another MAC can achieve all needed requirements.
Export Manager at ANHUI EMI TECHNOLOGY CO.,LTD ,manufacturer of AMR water meter,single jet ,mult jet ,smart water meter
5 年we are smart meter manufacturer,use all technology mentioned in this article,in this regard ,i think something right but not all
Senior Machine Learning Engineer at Faculty
5 年Intriguing. Like the Taiwan connection too...
Hexing Electrical Co,.Ltd - Manager of European Markets （区总）
5 年Very interesting all the comments here, good to hear and learn from different experts on different technologies.