Chain selection for a decentralized physical infrastructure network (DePIN)
Recently at onocoy, we had to make a chain selection for a decentralized physical infrastructure network (DePIN). onocoy is operating in the GNSS industry, that includes e.g. GPS. Since a) there are a lot of different chains to choose from, and b) DePIN is a novel use case for blockchain technology (alongside e.g. gaming, NFTs, and DeFi), we decided to share our methodology, final decision and lessons learned with the community so that others have an easier path ahead of them when it comes to chain selection.
The DePIN network
The purpose of this section is not to illustrate the DePIN network ‘onocoy’ in great details, but to give the reader enough information to understand decisions that have been made in the following sections (e.g., selection criteria).
Following the example of Helium, onocoy aims to build a dense network of RTK reference stations covering all land masses of the Earth (100,000 stations are required) to improve positioning obtained with GNSS (e.g. GPS) from meter-level accuracy to sub-centimeter accuracy. This would enhance existing applications (e.g., land surveying, autonomous vehicles, coastal navigation, etc.) to serve an established market (the medium-term goal is to have 2 million of these customers on board), but also enable new applications such as passive radars, identification of spoofing attacks, and advanced weather mapping.
onocoy is a collaborative project that solves a real tragedy of the commons in the GNSS industry: All the established players in the industry want a global and dense RTK reference station network, but the current incentives (e.g., cost, revenue, etc.) are set in such a way that this network is not being built. onocoy realigns these incentives by: Token-based rewards to cover investment costs in the infrastructure network through crowdsourcing, as well as transparent and inclusive governance to facilitate adoption and convergence of competing GNSS industry players on the onocoy platform. As these are the main use cases for blockchain technology, apart from the automation that is also required in this DePIN network, onocoy opted for building its system on top of an established smart contract platform. Notably, onocoy’s goal is to be fully decentralized in its final state, which means that all key interactions in the system will be enabled or secured via the blockchain.
The incumbent GNSS industry and its customers are not considered crypto-savvy and tend to be risk averse, so they expect the onocoy platform to provide a similar experience to what they are used to from web2. This is why a deflationary burn-mint model was chosen for onocoy, as it allows for stable fiat prices for end users that the GNSS industry is accustomed to, while allowing for strong token appreciation (as recently proved), which is important for the success of a blockchain project.
Chain selection methodology
We sampled by convenience a long list of potential chains which would facilitate the implementation of the envisioned DePIN system. We utilized a simple heuristic by starting from systems that were familiar to us and then extending this list whenever we identified a new system in our research. In total we considered 35 systems.
On this long list we applied a set of rejection criteria. As long as one of this criteria was applicable to a system it was removed from the list, leaving us with a short list of 8 systems. The rejection criteria and their reasoning are as follow:
- No smart contracts: To avoid being challenged by building its own blockchain network, onocoy opted for an established L1 blockchain on which systems can be implemented via smart contracts functionality. Hence, if a platform does not offer smart contracts, it was removed from the list.
- Multi-chain system: Out of complexity consideration, onocoy also decided against utilizing a multi-chain system which would still require from onocoy to maintain an own consensus/ dealing with additional cryptoeconomic complexity.
- Permissioned system: Trust and transparency are central to onocoy and therefore require a permission-free blockchain system. Therefore, permissioned systems were rejected.
- Small community: As community is everything in web3, systems with a small community have been removed.
- Small maturity: As onocoy is serving a risk-averse GNSS industry, systems with a small maturity/ age have been rejected.
The remaining 8 system were compared based on a set of selection criteria. These are
- Decentralization: Strong decentralization of the L1 is needed to fend off state-level attacks that could occur once onocoy is fully deployed due to the nature of the service it provides (accurate positioning).
- Scalability of the platform is necessary, as in the near future already 2 million users and 100k reference stations will regularly interact with onocoy and the underlying blockchain. A conservative estimate already results in 22 TPS (transactions per second) required by onocoy, more than many blockchains can handle.
- Low transaction fees are desired to make onocoy’s business model viable and not overly burdensome to customers who primarily initiate transactions in the chain.
- High market cap of a system as a proxy for its economic viability and thus likelihood of survival.
- On-chain governance: Possibility to participate in L1 configuration and development via voting as a further indicator of decentralization.
- Funding possibility in the ecosystem of the L1 — ideally commitments can be obtained before a chain is selected.
- Liquidity of the L1 token, which facilitates the tradability of the onocoy token.
Based on the previous comparison, 3 systems were collected that were than compared in detailed based on
- Technology (incl. TPS, blocktime, trx fees),
- Liquidity of native token in order for users of the platform to exchange onocoy tokens with third parties.
- Ecosystem including criteria such as amount of developers, existence of other DePIN networks on the chain, and the focus of the chain on DePIN networks. These are all important criteria for onocoy guaranteeing that it can benefit from network effects such as common/community tool development, shared marketing efforts/ conferences/ security audits, etc.
- Securing funding should be enabled/ made easy by choosing a chain (e.g. having a rich grant ecosystem, positive sentiment by investors, etc.).
- Risks of choosing a platform should be manageable (e.g., technology failures, ecosystem shrinkage, etc.).
Based on this detailed comparison, a chain is selected on which a detailed risk analysis is performed. If this risk analysis results in a chain rejection, another candidate of the top 3/8 should be chosen. In the case study illustrated in this article such iterations were not required.
Results from the case study
The following figure illustrates the 35 system and the selected systems after each refinement step of the utilized methodology.
Three system are identified to be well suited to implement the DePIN system under consideration: Algorand has a fair consensus and very secure/ stable network. Polygon is scalable and has access to the resources of the Ethereum ecosystem. Finally, Solana is scalable, has cheap transactions and a large ecosystem. These last reasons made onocoy select Solana as its L1 blockchain. In particular, for the following reasons Solana has been chosen:
- High throughput: Solana theoretically facilitates 65k transactions per second (TPS). At the time of writing, 4k TPS are observed of which 200–500 TPS are related to user activity. An all time high with 8453 TPS has been recorded in the past making it the smart contract platform with the highest ever observed TPS. A high throughput is important for onocoy, because in the medium term it is expected that it will have 100k reference stations and 2 Million users which will will result in a large amount of transactions initiated to the chain (once onocoy is fully decentralized). Assuming a conservative estimation of one interaction per user/ miner per day this already results in 22 TPS, more than some chains are currently (able to) handle. Solana has the capacity to cater for an increasing TPS need of onocoy.
- (Reliably) cheap transactions: At the time of writing the average cost per transaction is 0.00017 $ (0.0000079 SOL) on Solana. Assuming that for each miner and user (100k miners/ 2 million users) the blockchain has to be accessed once per day, this sums up to 364$ per day that is required to be paid by onocoy or its customers. For one of the other candidates the minimum requirement for transaction fees already results in 21609$ per day (59 times more) that would have to be paid. Also, other chains had spikes in transaction fees in the past that would have resulted in a cost at those days of >1 Million $ for the onocoy network paticipants. Such peaks in transaction fees have not been observed with Solana.
- Fastest block time: Solana has a blocktime of 0.4 seconds. When a user experiences more than a second of latency in an app (the brain processes information it sees at <0.15 seconds), it can feel like an eternity. With Solana onocoy is able to build DApps that allow seamless interactions for (non-crytpo native) users from the GNSS industry.
- Large developer base: Solana has the second largest developer base after Ethereum. A large developer base is important for onocoy because it can leverage the community to develop tools on top of its platform (e.g., as it has happened in Helium, e.g. transaction extractions tools, or novel types of dashboards).
- Focus on DePIN projects: Solana has communicated the goal to be the to-go platform for DePIN (e.g., their head of communications on Twitter). Hivemapper is already on Solana and Helium voted to move to Solana. This emerging DePIN ecosystem is of importance for onocoy, because synergies with these projects can be utilized/ explored (e.g. a tool has already been implemented that allows (non crypto native) users to pay with FIAT for service access within a deflationary burn-mint token model, which onocoy is also utilizing.)
- Clear roadmap to scalability, decentralization and resilience: Several initiatives and projects withing the Solana ecosystem work on further scaling and decentralizing it. E.g. a second validator client is currently in development which will improve the decentralization of Solana and potentially also scale the blockchain further. Moreover, programs are in place to decentralize validators. Also, updates such as QUIC and local fee markets were introduced to improve the resilience to DDOS attacks, also a decentralized indexer has been developed. A decentralized and resilient L1 blockchain is of importance for onocoy as it potentially can be under attack of colluding nation states.
- Popular programming language: Programs (smart contracts on Solana) can be written in Rust and C which are amongst the most popular programming languages worldwide. This will make it easier for non-native crypto developers from the GNSS industry to onboard on the code stack of onocoy. In particular, Solana has perfected the webdev to web3 onboarding pipeline.
In order to have a complete overview of the possible consequences of the selection of Solana for onocoy, a detailed risk analysis for the next 24 months has been conducted. The three main risks related to the selection of Solana as L1 are:
- Funding risk: Due to the FTX scandal, investors seem less likely to invest in Solana based projects. Nevertheless, despite all the negative reporting, VCs are still investing into Solana based projects (e.g. sec3), though we experienced concerns from some investors. Nevertheless, we also identified VCs that would invest in projects built on Solana.
Not receiving funds would make it necessary for a DePIN system to pivot either on technology, business model, legal structure or financing strategy, thus the materialization of this risk would have a large impact on onocoy. However, it is important to remember that the underlying chain is only one factor in the overall rationale for or against investing in the minds of investors and is not the first priority; things like numbers, brand, and community matter more. Changing the chain alone does not make these other points go away.
To mitigate the financing risk that might result from choosing Solana as the L1, a DePIN system should carefully elaborate their requirements for an ideal L1 and then illustrate how these can be met with Solana (see previous points 1–7). - Technology risk: Several outages were observed in Solana in the last years (e.g., due to DDoS attacks, bots, and bugs). The core technology of onocoy will be only slightly affected by such outages in the next two years. However, a decentralized system, as onocoy aims to be in the future, would be severely affected, preventing the issuance of rewards or user access to data streams.
To mitigate the impact in the near future, onocoy is developing its system such that it can enter information to the blockchain once it comes back online. In addition, the system is designed so that downtime does not negatively impact miner rewards or user costs. In addition, data streams are accessible to users even when the blockchain is down.
Furthermore, Solana mitigates this technology risk by introducing priority fees and identifying large transactions so that large token drops do not overload the network (local fee markets). In addition, the QUIC update to the data transfer protocol is expected to reduce Solana’s vulnerability to DDoS attacks, a major cause of previous network outages. In addition, a second validator client is currently being developed, which could reduce the risk of bugs, since the probability of them occurring simultaneously in two independently maintained code bases is low.
However, if these outages continue for more than 2 years, onocoy might have to pivot its chain. - Centralization risk: Centralized validators & foundation influence on Solana could result in a resilience risk to onocoy, if either a) malicious actors take over the consensus or b) the association stops to function halting maintenance/ development of the network. In particular, in risk scenario a) an attacker might change important entries for onocoy retrospectively in the blockchain or inhibit entries from being entered (“censored”) if having a majority in the consensus. Such an attack would have a large impact on a DePIN system as its customers/ stakeholders might lose trust. Solana has currently a Nakamoto Coefficient (NC) of 32 in its validators making it to one of the most decentralized systems in this metric. Requiring a higher NC would probably result in choosing Bitcoin or Ethereum, which come with other tradeoffs such as transaction fee cost. Solana is actively mitigating this risk further by decentralizing its validators via a) a second validator client and b) incentive programs for validator setup. Also cost for hardware required to participate in Solana consensus is expected to go down over time potentially resulting in a further decentralization of the consensus. In order to mitigate this risk, a DePIN network can implement an own warning system to identify such types of attacks and than have mechanisms in place to resolve them off-chain. As this is a major security flaw in the underlying infrastructure, a DePIN system would be required to move to another chain if this happens. Solana is also actively mitigating risk scenario b) by having its code open-sourced and nurturing a large developer ecosystem that would be able and willing to maintain the chain even if the foundation stops functioning. Moreover, the likelihood of this to happen in the medium term future (~2 years) is low as the Solana foundation is well funded.
Conclusion and lessons learned
Based on the above analysis, we have concluded that the benefits of Solana outweigh the risks many times over, and that Solana is therefore the right choice for onocoy’s DePIN. In particular, Solana promises to become the infrastructure layer that will facilitate the mass adoption of high-precision positioning by onocoy in the future.
While executing the chain selection methodology, three takeaways were identified that are briefly illustrated in the following.
- There exists (currently) no perfect chain that would check all desired requirements of a DePIN system (and probably never will). All come with some tradeoffs and have risks associated with certain components in their (eco)system (e.g., high security guarantees, such as those achieved with a proof-of-work consensus like Bitcoin uses, come for instance with a tradeoff in energy consumption). Researchers involved in the selection process have to be aware of this observation in their discussions. In particular, a leap of faith based on top of a rigorous analysis as illustrated in this article may be the right way forward when it comes to chain selection facilitating a right balance between rigorousness and efficiency.
- Obtaining funding from a chain’s ecosystem before a clear commitment to that chain is from our experience unrealistic. In particular, an ecosystem of a chain would first like to see that you are already building on their chain. Thus this cannot be a viable selection criteria (e.g. in order to identify the top 3 chains).
- Different modes of operandi in web3 and established industry require well-thought communication strategies. In particular, web3 people are often interested in a transparent and open process in which concerns are debated publicly, while established industry members often prefer a solid communication of results that provides security. Therefore, we recommend that other communities follow the method described above rigorously and transparently with those interested in or having an opinion on the chain selection, and share only the final selected chain with the broader community, along with the benefits and risks of that chain as presented in this article.
