Blockchain’s Role in a Decentralized Wireless Network

Despite the hype, blockchain is terrible for building applications due to its slow speed (pairing mining activity and transaction processing), loss of control (on a public, permissionless blockchain no single entity controls the blockchain), and increased complexity (including fuzzy business logic and running without trust).

In a post by our VP of Engineering, we wrote why organizations need to be very skeptical about using blockchain compared to a database.

However, despite these drawbacks, blockchain has proven itself as an effective underlying technology to build decentralized networks. In addition to enabling a reward system without middlemen, blockchain can provide an immutable audit trail to verify time and location of transactions as well as allow IoT devices, once onboarded and added to the blockchain, to connect to any hotspot without needing to re-authenticate.

We knew this type of decentralized wireless network would require a blockchain that could support the following capabilities:

• Fast transaction speed

• Ability to perform micropayments

• A new type of consensus model that doesn’t require computation

To date, many existing blockchains have been characterized by slow transaction speeds (typically 5 to 7 transactions/second), payment models that don’t support micropayments, and expensive offchain storage. And the most successful consensus models rely almost exclusively on heavy computation.

To build a decentralized wireless network, we’d need to build a new protocol blockchain and to truly scale, build two parallel subsystems: one to route sensor data to the cloud; and a second for the blockchain to process transactions asynchronously. For more details about our approach to this problem, go here.

Any decentralized system needs a way to establish trust in an inherently trustless way. Proof-of-work, while effective as a “sybil control mechanism,” requires heavy computational work for limited use, and creates a number of unintended side effects beyond slow transaction times. These include environmental damage due to power utilization, and formation of centralized mining pools which has made participation restrictive, unfair, and led to censorship of transactions.

To meet the needs of entirely different use cases that include collecting data from the physical world and adding it to a blockchain, we decided to build an entirely new type of proof. One that could establish trusted identity and instantiate a BFT consensus group based on our design goals and learnings from existing blockchains to avoid repeating unintended consequences.

We named it Proof-of-Coverage and key design goals included providing direct benefit to the network, being energy efficient, and maintaining fairness as the system grew. For details about Helium’s Proof-of-Coverage, see here.

Our implementation involves assessing nodes (we call hotspots which also act as wireless gateways) of the network on the quality of wireless coverage they provide and based on that quality hotspots earn a score.

By using peer hotspots to assess wireless quality, Proof-of-Coverage does not require computational power, and the scoring system ensures fairness by factoring network activity of a hotspot.

Pairing Proof-of-Coverage with a Byzantine Fault Tolerant protocol was our approach to complete the consensus model. However, we discovered to build this new type of decentralized wireless network we needed:

• No timing assumptions: assumes messages eventually get delivered

• Censorship resistance: miners cannot look into transactions prior to agreeing upon publishing

• Permissionless: Anyone can join the network

Thankfully Andrew Miller and his team from the University of Illinois heavily researched existing BFT protocols and came up with a new BFT called the Honey Badger. To learn more about our Honey Badger BFT implementation go here.

With this new useful, energy efficient, and fair consensus model we implemented an entirely new way approach to mining over radio frequency using Proof-of-Coverage with Honey Badger BFT:

• Hotspots on the network earn a score for providing wireless coverage. This is assessed and validated by their peers.

• The highest scoring hotspots are elected to consensus group.

• The consensus group performs the ‘work’ including publishing blocks, and receiving mining rewards.

• A new consensus group is elected and the process continues.

To build a network shared and owned by everyone means anyone should be able to participate. But to motivate individuals to participate, the right incentive model needs to be in place that doesn’t require a middleman.

An incentive model that can operate without any centralized authority is an area blockchain excels and has a track record of providing and delivering rewards using cryptography and advanced algorithms to replace a central trusted authority.

In a previous post, we described a new approach to build a wireless network for the IoT without telcos or other centralized organizations.

Connectivity to phones and computers is controlled by centralized organizations, and there’s a new race to control connectivity for everything else.

To join the race and participate in building this new decentralized wireless network, visit here.