The Challenge of Off-grid Connectivity
It’s no secret that mesh-grids are dependent on having internet connectivity to operate. Our Okra Pods need to sync billing data to the cloud so that people can use their energy, sensor data helps predict and detect issues automatically, asset performance and utilization data helps customers optimize their projects and much more.
Our in-house and open-sourced communications hardware is what makes this all possible. Today we’re unveiling the story behind our latest technological ‘hop’ that will prepare mesh-grids for their next level of scale.
The Approach & Challenges
Our Pods started with 2G cellular connections. Every single device was equipped with a 2G module and a sim card. That solution was low hanging fruit, as 2G networks are widely available and 2G modules are affordable and easy to program. It was the perfect starting point to bring our first devices online. Unfortunately, it was not the perfect solution at scale.
Off-grid areas often have poor to non-existent mobile network coverage, and the cost of running a SIM card in every single Pod has made our SaaS fees unaffordable. In addition, recently we have seen civil unrest lead to fuel shortages in places like Haiti, bringing diesel-powered cell towers down entirely. We are even looking into powering cell towers themselves from mesh-grids, an exciting future opportunity!
To address the 2G limitations and costs, we experimented with a satellite and WiFi model. A Very Small Aperture Terminal (VSAT) router was installed in a central location (powered by a Pod) and a WiFi network was distributed to all remaining households using repeaters. The Pods themselves connected to this WiFi network and operated normally. This solved some problems (removed the cost of SIM cards in every device), but created new challenges due to the severe range limitations of WiFi. Extending WiFi hundreds of meters or several kilometers throughout the community required a very high number of repeaters, increasing installation costs and complexity substantially.
We were determined to find an alternative solution that would address all our connectivity challenges and pave the way for mesh-grids to scale. This exploration led us initially to WiFi mesh and LoRaWAN, but further research identified those solutions as not suitable either, because of price, range or data rate limitations. But finally we found a protocol that looked promising: Wi-SUN.
Enter Wi-SUN
Wi-SUN (Wireless Smart Ubiquitous Networks) is a wireless communication standard that enables seamless connectivity between smart-grid and IoT devices. Fundamentally Wi-SUN is a standardized radio protocol that leverages license-free radio bands (868MHz in Africa/Europe, 915MHz in America) and enables meshing between nodes, meaning neighboring devices can forward data through other nodes until reaching the central gateway, the border router.
Similar to Zigbee or 6LoWPAN it uses IEEE 802.15.4 for its physical layer. It is widely used in on-grid electricity, gas and water metering - applications that align perfectly to our use case. Distance between devices can be up to 700m, which is suitable for 99% of the villages we intend to energize. With a data rate of 50 kbit/s, Wi-SUN supports a throughput for IoT devices sufficient for regular communication as well as over-the-air firmware updates. Our list of requirements were not easy to satisfy (simple to install, low cost, high data throughput, long range, high reliability) but Wi-SUN seemed to meet them all.
What will this look like in the field? We install a single VSAT, starlink, or cellular connection in a village as the uplink, and all the Pods in the village can send/receive their data via a Wi-SUN mesh network. Data from households on the edge of the village will “hop” through nearby houses until reaching the border router.
Engineering Implementation & Results
We started development back in May 2023 at our company-wide hackathon. Originally, the firmware development cost was estimated to be relatively high, because in contrast to 2G or WiFi, no off-the-shelf Wi-SUN modules existed. But the lack of a viable alternative protocol and severe importance of this work drove us to take it on. We developed hardware based on Texas Instruments CC1352P7 wireless MCU and wrote new firmware for that MCU as well as adapted existing firmware for our Pods. We have since published that new wireless MCU firmware on GitHub at: https://github.com/okrasolar/wisun_iot_modem
Once the firmware and hardware were ready, we deployed our first-ever field trial in a remote village in Cambodia. We installed a Wi-SUN border router (named Beacon) and retrofitted multiple 2G Pods with our newly-developed Wi-SUN modules. This trial gave us important data about actual field performance (almost nothing worked at first!), which led to multiple bug fixes, new antenna selection, and stability improvements. After learning from these issues, we deployed another trial village within an Okra staging setup in rural Australia, bumping up to twenty Wi-SUN Pods. This “village” continues to allow us to experiment with this new technology and gain confidence without interfering with our customers’ connectivity.
We’ve now had these Wi-SUN enabled Pods running stably for many months, a massively exciting outcome to prove this project and technology viable. Next up, we’ll be deploying our first Wi-SUN project with a customer, targeting August 2024, which will unlock full scale rollouts in 2025.
Conclusion
From the moment we built our 2G solution 6 years ago, we’ve always known we’d face a time where relying on 2G would hinder our growth and progress. And a year ago, with customers losing 30% of their revenue due to network outages, us making a loss on our SaaS offering, and our SaaS price becoming unsustainable, we made the strategic call to invest huge engineering resources into finally solving this problem.
Wi-SUN will enable us to solve all these issues simultaneously, and it has been built-in as a standard feature in our newest generation of Pods. It is truly a win-win solution, removing our biggest obstacles to scale while providing a better, cheaper offering to our customers. Additionally, site selection becomes much simpler, as we can effectively serve areas with patchy or non-existent cellular coverage, removing connectivity as a barrier to scale.
And lastly, zooming out to the technological landscape, our implementation will be the first application of Wi-SUN in an off-grid environment, a technical achievement we’re very proud of! This technology unlocks massive opportunities in bringing reliable, affordable services to remote parts of the world, and we’re thrilled to be pushing this innovation forward.
Stay tuned to see as we intend to roll out the first pilot of Wi-SUN built into our newest Pods in Nigeria in the coming months.
