Blockchain Commercial Solutions Pt.1 - Oracles
The First Money Lego: Bridging Real World Data and Blockchains
June 14, 2021
Enterprises will be forced to adapt to the Web 3.0 customer who expects fully reliable agreements cheaply. Why would a customer enter into an insurance policy with doubt around their claims being executed by the counterparty when they could more easily enter a policy executed by unbiased code? The latter is quicker to enter and exit, cheaper, and will consistently execute claims without counterparty risk. This narrative will play out across verticals. People don’t want to have to trust brands to do what they say they will do. Everything works fine until a crisis hits, then brands renege. We can look to AIG’s CDS obligations in 2009, the spectrum of insurance obligations when the COVID force majeure hit, or Robinhood’s recent suspension of certain tickers on its platform for examples. Further, in all of these cases, the ambit of the underlying risk was largely hidden from customers, only to become apparent when chaos had set in. Smart contracts propose an alternative: to reduce counterparty risk, bolster transparency, and execute without input from a human and all his foibles.
For smart contracts to become welded into commerce, they’ll need highly reliable real-world data to consume. We discussed the fundamentals and market snapshot of oracles at length in a previous post, so please refer to it if you’d like a refresher. Here we will concentrate on the commercial implications of these protocols. Oracles are meant to translate reality into a language that a smart contract can read. A properly structured parametric, or automatic, contract only becomes more efficient and more reliable than the incumbent digital contracts if it can competently verify that certain conditions are met. An oracle network acts as both the rails and the refinery for data, the internet commodity, that smart contracts consume. Oracles create the bridge between the real world and blockchains, condensing noisy data into actionable information. Oracle networks serve as a critical abstraction layer between meat space and commercial smart contracts—they ship truth to them.
Oracles: A Bedrock for Enterprises on Blockchains
When enterprises build out smart contract-enabled products, they’ll consider the following, among other things:
A) Data Fidelity: Sourcing as much real-world data as possible and refining that data into a form that customers believe in and that smart contacts can consume.
B) Secure Data Proofs: Proving that assets and high-fidelity data sets exist in a trustless and secure way.
C) Cross-Chain Compatibility: Integrating across incumbent blockchains (e.g. Ethereum, Solana, PolkaDot, ThorChain).
Data Fidelity
To the first consideration (A), blockchains are not inherently built to incorporate data external to the transactions occurring on-chain. Base layers are good at two things: 1. Transferring tokens and 2. Tracking signatures from agents verifying transactions. Blockchains are hyper-reliable within the confines of their walled gardens. To import data directly into a blockchain from a single source external to it would sully the continuity of the system, as consensus could not be achieved.
Smart contracts can verify certain conditions and execute value transfer if such conditions are met. They cannot inherently know things about the world. So an off-chain system needs to provide truth about what happened and feed the contract. Sergey Nazarov of Chainlink calls the chimera generated by merging a on-chain smart contract and an off-chain data source a hybrid smart contract.
An enterprise translating its digital agreements to smart contracts will generate little incremental value if the data feed remains centralized and subject to manipulation and shoddiness. This simply automates artifice. For example, imagine if AIG converts its insurance contracts to smart contracts but maintains control over the data that feeds the contracts. Sure, the claims process and its nuances are avoided. But the potential for exploit simply moves from the claims process to the data sourcing. Customers, particular institutional ones, will recognize this sleight of hand and elect for an insurance provider who uses decentralized oracle networks for data feeds. Less cynically, for customers to believe in the reliability of coded contracts, the data feeding them needs to be accurate. Effective oracle networks will leverage crypto-economic protocols to incentivize decentralized nodes to provide truth that smart contracts can ingest. A single data source may not be deterministic, but with the right incentives, a decentralized network of nodes can be.
Let’s explore how an oracle network might use crypto-economic mechanisms and incentives to deliver reliable, high-fidelity data to commercial applications:
Reputation: each node has its history of performance appended to it. A publicly visible reputation “score” will incentivize the submission of accurate data.
Majority Consensus: the range of outputs submitted by a given data network can be hemmed by excluding outlier data points and punishing the nodes that supplied them.
Virtuous Cycles: good faith submissions by data providers are rewarded. Bad faith submissions are slashed. A node is incentivized to provide good data for two reasons 1) immediate rewards for services rendered 2) the right to future rewards on the network
An oracle network’s ability to deliver higher-fidelity data lies in the embedded incentives. These well-structured incentives will support a step-wise increase in the quantity and quality of data fed to the smart contracts deployed by Fortune 500 companies and upstarts alike. Any institution or individual with a cryptocurrency wallet can supply data to the network and be rewarded or consume data from the network and pay for it. The reward and slash system pushes the average input quality higher and higher. The openness brings data quantity and the incentives bring data quality. This potent combination unlocks new frontiers for traditional industries like insurance.
We discussed decentralized insurance products in a prior Vulpine piece. In these cases, the oracle network will sample the real world for data—data that a smart contract compares to the conditional threshold required for it to execute. For a drought coverage contract to issue claims, for example, it will consume data from a network of rain gauges distributed throughout the coverage region. The oracle protocol will calculate the majority consensus input, rewarding those nodes that fall within it while slashing those nodes which do not, and submit the output to the smart contract program. When external data sources are properly incentivized to provide good data, insurance underwriters can excise their actuarial budgets and focus on providing a broader and better suite of products. And with a sufficient number of data nodes, the policyholders obtain comfort in the fidelity of the data ported on-chain. Once data is submitted, insurance contracts either execute and pay claims or do not. The process is more trustless and executes claims immediately.
As the value locked in smart contracts for a particular vertical grows, the market will require that the number of nodes and the value bonded by the nodes informing those contracts grow commensurately. i.e. A smart contract representing terrorism insurance on the Burj Khalifa in Dubai will be fed by more than three data nodes who each bond $1,000, respectively. This gives an obvious advantage to those oracle networks that move first and quickest. Prospective data nodes will look to join existing oracle networks because the demand for the product is highest and because the rewards are denominated in a token that has history and is worth something. Oracle networks will span the globe and verticals. All a company has to do is plug in and its enterprise solution is live. The data condensed through an oracle network is also more granular, so corporates can price things more tightly, increasing capital efficiency and lower working capital requirements.
Secure Data Proofs
To the second consideration (B), it is important to prove the existence and essence of an asset for agents to trustlessly exchange value over blockchains. Whether proving that a DeFi primitive (e.g. stablecoins, MakerDAO loan collateral pool) is genuinely backed by off-chain collateral or that a data set exists and contains signal (as opposed to noise), validity proofs are an instrumental building block for exchange. If an oracle network can confirm the economic potential of an asset base or data set, agents can tokenize these assets, imbuing them with property rights and “fixing” their economic potential. 1 By tokenizing data sets and verifying who owns, has sold, and has bought what, smart contracts expand the permutations of value exchange.
Let’s consider for a moment how companies train machine learning models. Typically, a company will generate a data set from proprietary sources or by querying lower quality public feeds. It will then feed this data into its machine learning algorithm and hope to generate insights that generate revenue from clients. Two organizations within the same vertical, say autonomous vehicles, might both be better off if they could train their algorithms on their peer’s data. Company A might have a deep trove of data, but are loath to share this data without a guarantee of remuneration at an efficient price. A prospective buyer of these data, Company B, is reluctant to unveil its algorithm and concomitantly the reason it seeks Company A’s data. The need for trust stymies a bustling market for data exchange, keeping value locked.
A hybrid-smart contract might be the antidote to such a deadlock, obviating the need for trust. Imagine if Company A and Company B codified their exchange of value in an oracle-fed smart contract. Company A submits its data set and Company B its algorithm to a decentralized oracle network. The oracle applies the algorithm to a cross section of the data and tests for signal. The oracle network assigns a rating to the data set. The rating would reflect the quality of Company A’s data in the context of Company B’s needs. The smart contract would be codified to execute payment at tiers that correspond to the usefulness of the data. Or to whatever pre-agreed execution the parties build in. Both parties have circumvented trust issues by ceding power to a non-human and unbiased agent. Both parties maintain privacy and are forced to act in good faith. Both parties walk away with the same or greater value than when they started. 2
Finally, in verifying that financial assets and their associated cash flows exist, oracle networks can bring productive real-world assets on-chain. Again, providing the bridge between off-chain and on-chain. For example, an oracle network can prove that an individual or company owns a property title deed and confirm with an audit that periodic rental payments do indeed flow into a bank account. By proving these two pieces of data, oracles enable property owners to issue security tokens entitling holders to a pro-rata share of rental cash flows. To fractionalize real assets is to unlock liquidity. This principle applies to all financial contracts. We can imagine vast securitization markets in insurance, real estate, and micro-loans blossoming.
Cross-Chain Compatibility
On the final consideration (C), it’ll be important for an enterprise to build cross-chain functionality, as chain maximalism will limit its universe of customers. Blockchains will specialize in what they’re good at, so certain business cases will migrate to certain chains. Companies spanning multiple verticals will thus need to be multi-chain. If a company’s clients are sprinkled across chains, the result is the same. To capture value in a multi-chain world, enterprises will either A) commit resources to multiple development teams who build tooling for multiple blockchains or B) plug in to existing infrastructure that underpins the universe of blockchains. From what I can see, the closest thing we have to this interoperable infrastructure layer is an oracle network.
Meat Space —> Cyber Space
Oracles will serve as a real world bridge, not simply from off-chain data to on-chain contracts, but from enterprises to blockchains. Oracles currently serve data to trading platforms (Dex/AMM), lending platforms, insuretech applications, NFT projects, and Layer 2 scaling solutions, among others. Oracle networks as a technology will give birth to completely novel industries.
The oracle meta-layer will have been built out over many years and inextricably weaved itself into the fabric of blockchain architecture. In such a prolific role, these oracle networks have learned how to serve a wide array of developer interests. Oracles enable organizations to use the inherent advantages of smart contracts to achieve their ends. In this way, an oracle is more than a data bridge. It will become an onboarding mechanism, a shepherd guiding companies from meat space into the ether.
Hernando de Soto: Mystery of Capital