As standalone systems, Blockchains do not have direct access to real-world data. It needs to be added. Smart contracts, because they're built on top of Blockchains, suffer from the same malaise.
The Internet, most software, and even humans have the same architecture. Information needs to be accessed, processed, and added before the memory banks/brains can use that information later. Information/Data is added to the standalone systems via mediums which are also called oracles - a callback to ancient Greek mythology.
With that brief introduction, let's dive right into the comparison of oracles provided by different providers - Chainlink, Kylin Network, DIA protocol, and our own, Umbrella Network.
We'll go over:
We use these rubrics to ensure the elimination of our bias towards our product.
There are over 10 oracle services in the market today, some live, some on the way. Each of them adds a unique feature to the mix and as with all things in life - there is never a one-size-fits-all scenario that works out in the long run.
The oracles we picked - Chainlink, Umbrella Network, Kylin Network, and DIA protocol, are sufficiently ready, backed by a decent development team, and most importantly - present different architectural features at a fundamental level that helps each of them act as a foil to the others.
We would also go over token details briefly as they could also help depict the market confidence in the project.
Decentralized oracles take different routes (architecture) to achieve the same, or at least, similar goals. Here, we will show you how each of these decentralized oracles was built and how it tries to make data access faster, cheaper, more reliable, and decentralized.
Chainlink is essentially a network of nodes to retrieve reliable real-world data via three types of smart contracts. To facilitate communication, Chainlink breaks the process of data aggregation into three parts.
The first step involves the creation of a Service Level Agreement (SLA) that specifies the required data. Chainlink uses the SLA to match with the oracles in its network that can retrieve the desired data, while the user locks in LINK tokens in the match making-contract.
This is followed by data reporting, where Oracles connect with the real world to obtain the data required by the SLA. The data is then processed and sent to the smart contracts on the Blockchain.
The final step involves data aggregation and validation via the aggregation contract. The validity of each data set is assessed, and a weighted result is sent to the user.
Apart from its network of Oracles, Chainlink also interacts with independent Oracles. This is done via the Chainlink Core and the Chainlink Adapter, where the Core routes the assignments to the Adapter, which in turn acts as a bridge between the node and external data.
As the first entrant to the market (by popular perception), Chainlink captures the pros as well as the cons. It has the most number of partnerships - but also suffers from an aging architecture.
With about 300 odd data pairs and the fact that it has a separate smart contract for each datapoint, makes it fairly expensive for validators - and creates the need to continually mint LINK tokens which practically dilutes the stake of the entire community as a whole.
Additionally, to create custom API requests, there is a requirement to whitelist the API first. Thus, there is a transaction (and the associated cost) to request custom data and another transaction to deliver the data.
Finally, Chainlink data does not include any proof of consensus which could be problematic given the fact that companies need to set up their own chainlink nodes which are effectively siloed.
This means that becoming a chainlink validator shall be out of the reach of most people - whether that's a good thing or bad, we'll leave it for you to decide.
Due to the high costs involved in the addition of data, the costs are passed on to those looking to acquire data - significant costs to the tune of one transaction on Ethereum. Due to network congestions on the Ethereum Blockchain, sometimes, the GAS fees have breached the $300 per transaction mark in 2021 alone. Imagine paying $300+ to find out that the price of 1USDT is $1.
Moreover, the prices are updated only once per day - which kind of works out when dealing in highly stable stablecoins but starts to show its age when you factor in volatile or not-so-liquid assets.
The Kylin Network is another decentralized oracle and is powered by Polkadot. It employs a novel Nominated Proof of Stake for consensus on a given dataset. This consensus involves validators, who run nodes for the creation and validation of blocks, and nominators, who nominate validators by staking their tokens.
The Kylin Oracle is a decentralized data feeding entity, built on the Substrate framework, and is connected to Polkadot as a Parachain.
Polkadot coupled with the Substrate framework ensure data integrity and network security via the arbitration mechanism.
Kylin does this via five entities that contribute to its workflow:
Kylin's data marketplace enables flow of real-world validated data sets to users via the feeders with the cost of access to be paid in $KYL. The marketplace acts like a one stop solution for all dApp developers, with the hosted server acting like the caching layer.
Kylin network inherits the benefits of the Polkadot-Substrate ecosystem that is similar to Eth2 - before the release of Eth2.
For example, on Kylin Network, all data validation can be done on-chain. However, this validation is what you'd call a binary process - similar to how legislators vote on legislative items. This means that validators do not give real values and that prevents median calculation from becoming a part of the equation
This makes the Kylin network swift and scalable, but also very dependent on the Polkadot development ecosystem and its future developments.
The Kylin network architecture of rewarding the oracle programmer with a portion of every payment makes it extremely lucrative for the core team and less so for those who come in after the gates open
Decentralized Information Asset (DIA) Protocol is an open-source Oracle platform that empowers the DeFi world with real-world data.
It is powered by DIA, the governance token of the platform also used to validate DIA's crowdsourced data feeds.
DIA functions via a four-layer architecture:
Let’s talk about each of these in brief-
Data Collection Layer: DIA's data collection layer comprises Exchange Scraper and Blockchain Scrapers that scrape data from various credible sources such as Binance and Kraken for exchanges, and BTC and ETH for blockchain.
Data Storage Layer: The data storage layer is composed of a Time Series Database that optimizes the storage of time-stamped data, such as crypto-feeds. The includes the data verified by the validators on the network and stored on an immutable, open source database made available on DIA's platform. The primary function of this layer is to facilitate trades and circulate supply data.
Data Filter Layer: This layer includes two filters, the
Price filter, and the
Supply filter, with each exhibiting a unique role. While the Price filters perform a technical analysis of the price and volume of data, the Supply filters monitor and regulate the supply and circulation of the existing tokens and events.
Data Connectivity Layer: This layer is responsible for connecting dApps and other applications that require data and the data providers via Oracles or Rest API. While DIA protocol offers historical financial data for free, tracking the live price of any financial asset/cryptocurrency requires DIA tokens.
On the DIA protocol, to update a value, each validator has to submit a new transaction for every data point and the contract owner updates the price without validation.
On a technical level, the DIA protocol has only 40 lines of code with no dependencies and no libraries - could be a worrying feature or maybe it is intended to solely serve as a kernel.
Even with that benefit of the doubt, it is hard to think of the DIA protocol as anything more than a centralized server that scrapes data and posts it on the Blockchain without validation.
Think of DIA protocol as the Linux of decentralized oracles - secure, massive, and sometimes unwieldy. Potentially any data on any cryptocurrency across every blockchain is supported - but that can be said of other oracles too. Untimely audits and lack of developer support outside of mostly hobbyists throws a little dark cloud over this otherwise wonderful architecture
Umbrella Network is the world's first truly decentralized, community-owned, and highly scalable Oracle solution. It employs a DPoS consensus algorithm and offers a high-speed and comprehensive data feed of about 1200 crypto prices (with the stated goal of reaching 10,000+ prices by the end of the year).
The data relayed by Umbrella Network is divided into Layer-1 and Layer-2 data, with the latter being the sidechain of the main chain. Layer-1 relays the First Class Data (FCD) and comprises cryptographic proofs for pricing data. While the FCD includes the most popular crypto pairs along with some crypto-fiat pairs, the cryptographic proofs can be leveraged by users to verify the pricing data.
First Class Data: This data set includes some of the most popular and in-demand crypto pairs, crypto-fiat pairs, and implied volatility for BTC, and ETH is always available on the chain. This data Umbrella Network makes available free of cost, and most Oracles in the market deal in the same.
Layer-2 Data: This is a rich dataset comprising of crypto pairs, indexes, volatility data, and much more, and includes over 1200 data points. The Layer-2 data is stored in Merkle Trees, an efficient and secure data structure that anchors it to the Layer-1 data.
The Layer-2 data is stored and relayed by Umbrella Network using Merkle Trees. Also called a binary hash tree, Merkle Trees allow seamless transfer of data among nodes in a peer-to-peer network. Each leaf node represents a data point on a Merkle Tree and is rolled up to the root hash. The Proof of Stake also includes the root hash- representing the entire set of data points contained within the Merkle Tree.
What sets the Umbrella Network architecture apart from the others is:
This leads to massive savings on GAS costs which are leveraged in three different ways:
In a marked departure from the centralizing tendencies shown by other oracles that require a high upfront cost to run a node, Umbrella Network community validators can run a node with as little as $2k per month - even this cost is reimbursed. Moreover, Umbrella Network shall support multiple Blockchains and that foments greater decentralization.
What sets us apart from the other oracle architectures mentioned earlier can be summed up in the following bullet points:
Iron-clad Data Validation - Umbrella Network's smart contract verifies the addresses of all validators and consensus over data is sought (and achieved) before commitment to the Blockchain. The proof of consensus is attached to every piece of data - ensuring data integrity even after retrieval from the Blockchain
Single Architectural Framework- Some of the oracles mentioned earlier had different networks for different types of data. This creates siloes and runs the risk of needing separate transactions for inter-network data despite being on the same platform. Umbrella Network's batching of data and merkle-ization is the key differentiator in this regard - allowing us to bring our vast (largest amongst all compared) data points into the same network
Cascading Benefits of Fixed-Supply Tokenomics- Simply put, an unlimited supply of tokens basically slashes the value of HODLers every time new tokens are minted. In the world of oracles, that essentially means that first, you pay for the data directly and then you pay via value decimation indirectly. On the Umbrella Network, the supply is fixed and the token is used not to decrease the value of tokens held by HODLers but is used for DPoS consensus.
That oracles shall be instrumental in enabling cross-chain support, data access, and data retrieval is undeniable. It was on this note that we decided to compare the architectural strengths, weaknesses, and potentials of the four kinds of technical architectures used by decentralized oracles of the day.
Chainlink is the obvious leader as things stand today, with several hundred confirmed partners but questions regarding high costs involved need to be addressed. Nobody wants an Internet Explorer in a day and age when Chrome exists, in terms of analogy. We wish the best to the first leader of the industry.
Kylin Network is a great response to the diversification of Blockchain architectures and presents an excellent solution for at least the Polkadot ecosystem. However, it is still in the testnet phase (as of writing) and shall need to be battle-tested in the days to come and prove its mettle.
DIA protocol has a very modular approach to the entire Oracle ecosystem. Decentralized to the extent of a DAO on day 1 has its pros but that there is a long road to decentralization, as exhibited amply by Bitcoin, as well as, Ethereum should have tempered fiction towards reality. We wait on this front too, to see it in action, before judgment can be pronounced.
Finally, Umbrella Network, our answer to all of the above, already LIVE on Binance Smartchain, going live on Ethereum in September 2021, looks like the most promising to the Chainlink juggernaut. By actually offering prices at a fraction of what is normally acceptable and by onboarding its first cohort of validators, Umbrella Network decentralized oracles are diametrically opposite to several architectural features of Chainlink.
As DeFi grows and absorbs more talented developers, data-richness, affordability, and the highest level of freedom/decentralization shall determine who comes out on top. In our (objective but subjective arguably) opinion, we've ticked those boxes with our decentralized oracles. But, as goes the adage, it shall be seen.