A blockchain is a fully deterministic system meaning that, given the history of the system, it can be replayed in another machine achieving the same results. Essentially it is a machine that, given the same inputs, will always provide the same outputs.
This is a very powerful property, because it allows users to store information of different nature safely, with a very limited risk of data tampering.
This powerful property comes at a cost: isolation. Blockchains are isolated networks that cannot communicate with the external world without a layer of “translation”. This role is fulfilled by what is called an Oracle.
An oracle is nothing but a node in the network that acts as a bridge between onchain and offchain data, so a part of the system that is able to translate external pieces of information into a blockchain-comprehensible language.
External data is indispensable to build useful applications. So if the goal is to use blockchain for real world applications, oracles are key.
The goal of this post is to explore how the oracle space is growing and, specifically, what are the most interesting oracle projects in the DeFi real world asset space today.
Chainlink: the go-to oracle-as-a-service provider
Projects can build their own oracle solution or use dedicated infrastructure providers.
In the decision making process, it is key to consider that the data provided by oracles is very sensitive as it represents the blockchains’ interface to the rest of the world. If an attacker were able to manipulate the oracle price, the results could be terrible for a protocol or an onchain application.
Given its structural criticalities, today the vast majority of projects avoid building an oracle system themselves and move towards using oracle-as-a-service providers.
In this space one service has already emerged as the market winner: Chainlink.
Chainlink is not the only application that allows the creation of oracles (other relevant projects are Band Protocol, API3, Nest protocol) but it is by far the most used solution in DeFi.
A key characteristic that Chainlink provides is the decentralised nature of its oracles. Chainlink doesn’t take a piece of data and simply export it onchain, but it introduces a number of decentralisation layers that improve the resilience of the system and makes potential attacks more expensive, thus more complex to execute.
Specifically it has 3 layers of data aggregation:
- Data sources – Chainlink usually doesn’t fetch data from a single source of truth, but it differentiates it from multiple available sources and considers the weighted volume of all the data sources removing outliers, fake volumes and constantly monitoring downtime of each sources
- DON – Decentralised oracle network. The network is constituted by various nodes, all run by reputable institutions. Each node reads multiple data sources and calculates the median of the data points exposed by each data source.
- Chainlink aggregation – ultimately Chainlink synthesizes the various pieces of information coming from multiple nodes, into a single value. This is done by calculating the median of the data points propagated by the nodes.
In the end the final data point exposed by Chainlink is a median of multiple medians, all calculated pulling data from a significant number of data sources.
This data point is now frozen and made available onchain to data consumers, either in a predetermined reference contract or as a response to a previous request.
The Chainlink network (which as seen above is key to guarantee a decentralised system) is run by third party validators that do the job of fetching data in exchange for a fee.
Currently Chainlink provides a certain number of data feeds but, given its design, any sort of data can be brought onchain: weather, sport events, political events, market data.
Essentially Chainlink has positioned itself as a standard piece of infrastructure to build data application, not very differently from what AWS is for Web2 computing. Through Chainlink as long as a data point is exposed by a data source, it can be connected via an external adapter to the Chainlink Oracle network and then used by a smart contract onchain.
Real World data providers
If DeFi is the future rail of financial services, it is absolutely necessary to build a constellation of data services that can provide reliable data of real world assets and phenomena onchain.
For example, for mortgages or invoice financing applications to move onchain it is absolutely necessary to have some services that provide real estate and credit scoring data onchain. The list of key data points not yet published onchain is infinite but the ecosystem is growing very very fast.
The first vertical of data applications that is being migrated to blockchains is traditional financial market data: Equities, Bonds, Commodities.
Various serious players are now emerging in the space, providing censorship-resistant, decentralised data feeds.
The most used one is once again Chainlink. Chainlink Price Feeds, leverages the open source Chainlink technology (presented above) to create price feeds of the most important financial securities and make them available to onchain application for a fee.
A strong competitor to Chainlink Price Feed is Pyth Network. Pyth Network positions itself as an oracle that publishes financial market data, like US equities and commodities, to multiple blockchain. The chain supported are Solana (Pyth native chain) and various EVM compatible chains (Ethereum, BNB, Avalanche). Pyth guarantees the robustness and accuracy of its price feeds through a dedicated chain Pythnet, running on Solana technologies, and a set of oracles that runs on the Pythnet and execute various tasks: maintaining the set of price feeds; storing the contributions of data providers to each price feed; combining the individual data providers’ prices into a single aggregate price and confidence interval; performing any additional stateful computations on the resulting price series.
Overall, Pyth network seems a reliable solution with a stronger institutional vibe but less open source ethos than Chainlink.
In the very short term, the killer application for data providers will be bringing traditional securities data onchain. Nonetheless, I do believe that in the medium term different verticals will become necessary and in this direction, various projects are emerging. The two most intriguing ones are: dClimate and Truflation.
dClimate is a marketplace for climate data sources. It uses Chainlink to retrieve weather data from qualified data sources and validate it, adding some layer of intelligence and filtering.
The product is already used by onchain insurance projects like Arbol, to offer parametric insurance products: insurance products linked and triggered by publicly verifiable climate and weather data sources instead of human adjusted valuation.
Truflation is an inflation feed service that is trying to bring inflation calculation into the 21th century. They basically deconstructed and re-engineered the way inflation is calculated by governments, substituting the survey approach used by sovereign states with the real-time measuring of the online prices of a predefined basket of items. Their methodology is transparent to users and they provide inflation updates on a daily basis.
The use cases that could come out of this are basically infinite, given the centrality of inflation in our economic life, but the most interesting one in my opinion, is the creation of inflation-resistant stablecoins (Nuon by Laguna Labs).
The interesting aspect of both dClimate and Truflation is that they are both data providers that use the oracle-as-a-service module offered by Chainlink to make their data feeds censorship-resistant, decentralised and available onchain.
I firmly believe that oracles and the migration of data onchain is at the first stage of its evolution. My expectation is that in the future, millions of data feeds will be available to developers and, at that point, building a data feed will be orders of magnitudes simpler than today – just like creating a website used to be versus what it is today.
Chainlink is strongly positioned to become the industry standard for data feed operations – the AWS of oracles – but this doesn’t exclude the opportunity of creating solid projects dominating new verticals.
In this future world, the key problem to solve will be the lack of a solid vetting and curating system to select the most appropriate data source for your needs. If discoverability on market.link is already complex today with 500 data feeds, it will become impossible when this number will explode.
In the long run we are all deadJ. M. Keynes
Anyway, we are still not there, applications built today have to solve real problems of the existing crypto ecosystem to survive and these are coming from DeFi applications that are bridging real world assets onchain. Most of the DeFi excitement sits in this space today, and this will be the topic of my next post: All you need is Bond.
- Ethereum on oracles – Link
- A comprehensive view of oracles by Fabric Ventures – Link
- Why contracts can’t make API calls
- dClimate introduction – Link
- Pyth Network – Link
- ChainLink – Link
- ChainLinkGod podcast on ChainLink Products – Link1
- ChainLink Proof of Reserves – Link
- ChainLinkGod podcast + Truflation – Link