(This article has been written jointly with Adnan Warsi)
We are still in the early days of Web3. Web3 is expected to further revolutionize communication as we know it by making an individual citizen more powerful with tools of decentralization.
Yet, the challenge of bridging the divide between the physical and the digital worlds is substantial. In the use cases of purely financial transactions, the divide is relatively easy to bridge. As a result, we've seen an explosion of web3 in the world of finance in the form of Decentralized Finance aka DeFi.
There are several other areas other than pure finance that will be revolutionized by web3 going forward. One of these areas is wireless communication. That is our focus in this article.
Wireless Communication and Web3
Wireless communication is a domain where web3 has the potential to make a significant difference. Bridging the divide between the physical and the digital worlds is easier here than in case of other industries.
Most transactions in this space usually involve payments for transport of messages over the network. Both the transport of messages and the corresponding payments can be tracked in the digital domain entirely. This reduces the challenge of bridging information from the physical world.
As a result, we are optimistic about the vision of setting up a decentralized wireless communication network. In such a network, anyone can deploy a hotspot to provide wireless coverage. Devices can connect to the hotspot and pay to have their data transported to the desired destination. Setting this up as a DAO (Decentralized autonomous organization), while tracking payments on the blockchain, can provide the right incentives to all the parties involved.
There is already a commercial program underway towards setting up a decentralized wireless network in the market. Helium (https://www.helium.com/ ), a company that seeks to create a decentralized wireless communication network is focused on IOT communication at present. They may evolve to offer 5G connectivity in the future. Helium has their own token and their own blockchain to track transactions in the network. The approach they are taking seems to be a partial web3 approach (using tokens and blockchains only) at this time which is understandable given that Helium has been around since 2013.
In the rest of the article, we will focus on IOT networks. Extending this to cellular networks may appear easy since cellular networks are wireless networks too. Yet it can be a hard problem to design for since cellular networks entail issues related to spectrum licensing and regulations that differ across markets.
Wireless Networks
In this section, we look at the architecture of a wireless network. A wireless network will focus on ensuring that data can be transported from end devices to the cloud and back. The end devices contain sensors, data from which need to be transported to the cloud, so that the data can be analyzed, and decisions made based on the measurements.
At a high level, there are three types of components in the system:
End devices (with sensors): These are the originators of sensor data to be ‘sent’.
Hotspots: These provide the connection to the end devices and transport data from the sensors to the desired destination.
Cloud infrastructure: This includes the destination of each sensor along with other servers needed for the operation of the system.
An end-device connects to a hotspot and sends data to the cloud via the hotspot, as shown in the figure below.
Traditional approach
In the past, all these actors have been under the control of commercial entities. For example:
With cellular networks, a company such as Verizon or ATT controls the gateways. The end devices are the cell phones that belong to consumers. The cloud infrastructure may be under the ownership of the providers of the apps you use. This type is a 3-party network.
Or a single company (such as GroGuru) may control all the three actors. This is a single party network. Note that the single company would still leverage other commercial organizations such as AWS or Verizon, but that may not be visible to their customers.
The risk with existing approaches is the dependency on the organization that controls the gateways and cloud infrastructure, especially if coverage is desired in regions where it did not exist.
To address this, some organizations like ComCast have allowed third parties to deploy their own gateways. These third parties are still managed by a single organization. However, this has typically not been economical for any of the parties involved.
In addition, billing at an atomic level where each message sent is billed separately has not been practical in the past. But web3 makes billing at an atomic level feasible by reducing the cost of reliably tracking each transaction.
Both these features possible in the web3 approach make it possible for wireless coverage to be offered by a DAO. This is indeed what we will look at next.
The IOT DAO approach
We will first explain at a high level how this problem can be solved using the DAO approach. Having explained the DAO approach, we will then look at the Helium approach briefly.
The steps involved in setting up the IOT DAO then are:
A high-level design description of the IOT DAO is given below. Several open questions still exist. We plan to do this in the future.
Set up a DAO with its own token and a predetermined number of tokens in its treasury.
Parties that want to be part of the wireless ecosystem join the DAO.
Set up a blockchain to track the transactions in the network.
Miners are rewarded for maintaining the blockchain.
Some of the DAO members deploy hotspots to transport IOT data to the cloud.
The hotspots can initially be deployed on a “deploy when you want” basis.
The hotspots are expected to purchase and maintain their own backhaul.
Allow edge devices (with sensor nodes) to connect with hotspots and transport data using the hotspots.
To do this, the edge device will have to purchase tokens from the network.
These tokens are to be used to authorize the data transfer by the hotspot to the network.
The token transfer transaction will be captured on the blockchain.
Data transport transaction
The hotspots will get an authorization token from an edge device.
The authorization token will contain information such as
Number of tokens to transfer from edge device to hotspot for the transport service. The number depends on the number of messages, type of data delivery (reliable -TCP vs unreliable - UDP)
Long-lived sessions with messages sent over different time periods would require a simple proof to be included with each message authorizing the message.
The hotspot will forward the authorization token along with the response from the destination (for reliable delivery) as a signed message to a miner. This is considered a transaction and when included in the blockchain, will cause a transfer of tokens from the edge device to the hotspot. A portion of these transactions might also be transferred to the network.
As a result, the hotspot will transfer the message to the destination as a TCP or a UDP packet. The contents of the packet are not pertinent to this transaction.
Incentivize miners to start creating blocks of the blockchain. This need not be a POW chain. It can be a chain such as the PoX chain leveraging another crypto token such as Stacks (our preference).
The miners are awarded block rewards in the form of the network tokens.
We could also look at not creating any tokens initially in the treasury. Rather, tokens get created only when blocks are mined.
So in that situation the edge devices will have to purchase tokens from the miners instead of from the DAO treasury.
We will also need to set up a way to purchase tokens using fiat or other crypto tokens.
There are still several questions related to the block dynamics, consensus type, token type etc. We will have to decid these based on the traffic expectations on the network and other factors. This is planned for future work.
Economics of the DAO approach
The DAO needs to ensure that the money it makes justifies the tokens owned by the DAO. The DAO tokens should not create money out of thin air, so to speak. This is the basic principle that we use in evaluating the wireless DAO.
We assume that the economics for the wireless DAO should be similar to the economics for Helium. Hence, we have used the cost and price figures associated with the Helium network for our analysis here.
The DAO makes money from the customers who use the data transport service. These customers are the end devices.
We make the following assumptions in our model:
Each end device pays for data transport service via tokens; but these tokens correspond to fiat. Hence, for our modeling, we assume a typical device sends 24 bytes every five minutes, paying the equivalent of $1.05 over a year to the DAO.
80% of the revenue goes to the hotspot owner while the remaining 20% goes to the DAO as a fee. As the adoption increases, the DAO fee as a percentage of hotspot revenue will probably be reduced.
The miners get rewarded for mining each block. We ignore this and the transaction fees for our models for simplicity.
We assume each hotspot has a range of about half a mile. Because of this, it covers approximately an area of 1 square mile. Note that for a hotspot to get a range of up to 10 miles, as claimed by the LORA hotspots on Helium, the hotspots have to be mounted quite high. In our previous lives, we have seen that we will need to mount the hotspot at over 150 feet in order to get these long ranges. The implicit assumption here is that the hotspot provider deploys the hotspots in buildings which could be about 10 to 20 feet high.
We assume that each hotspot costs approx $500 to deploy and has a life of 5 years.
We also assume that the backhaul costs at each hotspot are $10/month.
At an installed base of 37 million LPWAN devices in North America in 2021, we assume that they’re deployed in regions proportional to population density. For a mid-sized city like San Diego with an urban area of 370 sq. miles, we assume an install-base of 200K LPWAN devices. For an evenly distributed density of hotspots within a mile , each hotspot will cater to just over 500 devices. At a projected CAGR or 35% over five years, the install base will increase 4 fold by 2026.
With these assumptions, a hotspot will earn $42 per month, of which 20% shall go to the DAO and a $10 monthly cost for backhaul fees. Over a lifetime of a hotspot as an investment, this translates to a 23% annual rate of return. If we assume that most of the hotspots will not incur a marginal cost for backhaul, as they’d typically use an existing on-premise Internet connection, the rate of return rises to 30%. In 5 years, with a 4-fold increase in devices, the effective hotspot effective rate of return on investment approaches 60%.
At the city level, this is equivalent to a cumulative revenue of $200K over a year, of which $40K goes to the DAO. We also have the view that with a robust economic DAO model, participation and adoption will be accelerated to be higher than typical 35% CAGR of most industry analysts.
A graph of the effective rate of return (RoR) as a function of number of devices on each hotspot with a 20% revenue split with the network is shown in the figure below.
There are several other questions that need to be addressed in case the reader is interested in a detailed report. Reach out to us if this is of interest. A sample set of questions are
If we are dealing with ISM bands (as Helium is), how do we ensure that a flood of hotspots does not cause a deterioration of the service for both the hotspot providers and for the end devices
How do we incentivize end nodes to switch to this network
How do we incentivize the expansion of coverage
Having described the DAO approach towards addressing the problem of providing wireless coverage, we next look at how Helium is addressing this problem.
The Helium approach
Helium is also focused on addressing the problem of coverage. Just like in the DAO setup described above, they propose the creation of a decentralized network where no single organization controls the entire network. Any third party can deploy a gateway to transport data from the devices to the cloud. The devices themselves pay for the transport of the traffic. In addition, Helium has its own token and blockchain. And all these transactions settle on the Helium blockchain.
The Helium network also consists of three actors
End devices are the devices that have data to send
Hotspots providing the link that end devices can use to send data into the network
The cloud consists of miners of the blockchain and routers which are the destination points for data sent by an end device
The network works as follows:
Devices pay to send and receive data from the network.
Miners own hotspots and are provided network tokens for providing network coverage.
Miners also get the transaction fees and the rewards for maintaining the validity of the Helium blockchain network.
The Helium blockchain uses a new consensus protocol - Helium Consensus protocol
The network also uses a new kind of proof called the Proof of Coverage - this is about a hotspot being able to conclusively prove that it is able to provide coverage in the lat/long that it claims to be in.
The miners can specify the price of data transport from hotspots to the routers and routers can specify what they are willing to pay.
Any miner can stake the HNT token to become a validator. And a small group called the consensus group is chosen randomly from the group of validators.
Validators get paid for maintaining the blockchain
a consensus group among the validators gets chosen to decide the next blocks of blockchain and they get the block rewards
the consensus group is randomly chosen from the group of validators
anyone can become a validator by staking 10,000 HNT tokens currently
The consensus group is responsible for creating blocks of the blockchain.
The consensus group is elected frequently.
We next look at the economics of the Helium network briefly.
Hotspot operators
They pay approximately $500 to deploy a hotspot.
In addition, they have to bear the cost for backhaul connectivity
They get paid for transporting data from end devices
Traditional Helium hotspots would transport data and also mine for HNT (this is just computing and sending proof of coverage to get HNT) while new data only hotspots forgo mining (the proof of coverage just seems to add to complexity is my opinion)
End devices
100 devices sending 24 bytes of data every hour costs approx $0.71 per month as shown in the image below.
Validators
Currently, with approximately 3,000 validators and 43 in the consensus group, each validator gets a return of 6.3% annually as shown in the image below.
This does not take into consideration that the HNT stake is locked for 5 months after the validator wants stops staking.
So the effective return is reduced to say 4.5% annually.
And this will go down as the number of validators increase as shown in the image below.
Data credits and burning HNT
Customer burn HNT to get data credits which allows the end devices to send data on the network.
Comments:
We believe that the Helium network has added a lot of complexity, perhaps attributable to the history behind the network.
For example, consider proof of coverage. Asking a hotspot to verify proof of coverage is similar to asking an Uber driver to prove his location. The focus should be on providing the service. You need the Uber driver’s location to ensure he gets paid for providing service. If the driver provides a false location, the driver will not be paid since the driver cannot provide the service. It is the same with a hotspot - if a hotspot provides a false location, then the hotspot will not get to transport messages and hence will not be paid.
Creating a separate token for the Helium network - the HNT token along with another token to be used for data transport adds friction to the network.
Having devices specify their routers on the blockchain is also not needed. Alternatively, the design can just have the routers purchase tokens to use for data transfer. And the data will be transferred to any IP address that the device wants to transfer the data to.
Summary
We have looked at how the problem of wireless coverage can be addressed as a DAO. We have looked at the basic design and a viable business strategy. While the initial study shows promise in this approach, detailed investigations will be needed before getting such a wireless network deployed.
References
https://explorer.helium.com/tools
https://shop.bobcatminer.com/products/bobcat-miner-300?variant=39570357911709
https://www.lightreading.com/open-ran/helium-aims-to-be-largest-cellular-network-in-us/d/d-id/774304
https://github.com/helium/HIP
https://www.coindesk.com/business/2021/04/27/helium-to-launch-5g-network-with-blockchain-powered-mesh-of-diy-telco-hubs/
https://www.eetimes.com/in-the-air-tonight-heliums-5g-dreams/#
https://www.helium.com/about
https://staceyoniot.com/helium-built-a-network-and-now-it-has-customers/
https://staceyoniot.com/how-i-made-10000-providing-part-of-heliums-iot-network/
https://www.lightreading.com/aiautomation/dish-to-use-heliums-diy-5g-network/a/d-id/773023