Band Protocol and Chainlink: A Comparative Analysis
Band Protocol is an oracle platform designed to aggregate and connect real-world data and APIs to smart contracts. Band v1 originally launched on Ethereum in September 2019 using a bonding curve design, but received little interest from developers and ultimately achieved no adoption on mainnet. It was abandoned less than a year later by the core development team for Band v2.
Band v2, referred to as BandChain, is an oracle network running on its own Tendermint based blockchain (using the Cosmos SDK). Band has launched two versions of BandChain, with both being named Guanyu. The first network is a Proof of Authority network with seven nodes belonging to the Band Foundation and private select validators, while the second network is a Delegated Proof of Stake network.
In this article, we will first explore how BandChain works from a high level in order to lay the foundation for a comparative analysis that details how it stacks up against Chainlink. The main focuses of the comparison are as follows:
Data Quality
- Band: BandChain only initially supports free, low-quality APIs, whereas paid, password-protected API support is still largely theoretical. Thus, developers cannot just call any password-protected API they desire. Their proposed plan is to require data providers to change their entire business model to accept on-chain cryptocurrency payments, drastically limiting the data available to users.
- Chainlink: Natively supports connection to any data source API live in production today, including free open APIs, paid authenticated APIs, and proprietary private APIs. This is all made available through modular external adapters that can be created and hosted by anyone without requiring any additional support from the Chainlink team or data providers themselves.
Reliance on Randomness
- Band: Built randomness as a core component of the protocol; it’s the required method for choosing oracle nodes, despite the feature not being supported by any original academic research. Not only does it severely limit data accessibility (every node must have access to the same data) and network security (must trust every node), but the team is largely incompetent on the subject, as they were caught trying to copy Chainlink VRF on their Github repo.
- Chainlink: Does not include randomness as a core part of the protocol, ensuring users have high levels of flexibility and quality control mechanisms for choosing nodes and connecting to data. They built Chainlink Verifiable Randomness Function (VRF) as a novel approach to creating a provably fair on-chain source of randomness for blockchain gaming and NFT dApps, backed by leading and original academic research.
Adoption Metrics
- Band: Minimal usage in-production on mainnet despite Band v1 being live on Ethereum for nearly a year. Band v2 lacks core features required by users, does not have a network effect, and cannot offer any case studies showing its ability to secure a large amount of value locked in the DeFi economy.
- Chainlink: Currently secures tens of billions of dollars in value on mainnet for top ranking DeFi projects through its numerous live Price Reference Feeds. It’s the most used oracle throughout DeFi, across blockchains, within gaming thanks to its pioneering work on VRF, and has a 400+ project integration pipeline.
Resource Efficiency and Network Design
- Band: Their oracle is built on its own blockchain, where oracle nodes are required to perform secondary jobs of being block producers/validators of BandChain, which adds no value to the oracle’s job of data delivery. As such, nodes experience high operating costs due to growing state bloat and long sync times as time progresses.
- Chainlink: Chainlink is not, nor will it ever be, a blockchain, so nodes only operate as oracles solely focused on data delivery and not block production/validation. Nodes are extremely lightweight and require no state processing or synchronization, allowing nodes to be spun up and utilized instantly.
Team
- Band: The Band team is minimal in size, hasn’t produced any original academic research, and lacks any experience building distributed systems and oracles, especially ones that secure real value in-production. They have also abandoned their original design and changed their protocol numerous times.
- Chainlink: The 40+ Chainlink team has been building oracles and presenting original academic research since 2014, and have support from top advisors like Ari Juels (former Chief Scientist of RSA), Tom Gonser (founder of DocuSign), Evan Cheng (Director of Engineering at Facebook), and more. They have stuck to their original design and expanded upon it to improve the experience for users.
Decentralization
- Band: All data queries/responses must process through a single blockchain, BandChain, limiting the maximum throughput. Has an upper bound of 100 nodes due to the limited scalability of Tendermint BFT consensus.
- Chainlink: An unlimited number of oracle networks can operate in parallel and run natively on any blockchain, DLT, or layer 2 solution, eliminating the need to funnel data queries/responses through any one chain. No upper bound on nodes with the coming threshold signatures and off-chain aggregation upgrades.
Multi-Chain Support
- Band: Relies on yet to be released Cosmos IBC to natively bridge data to other blockchains. Thus, it currently only supports lite clients, which require data to be delivered by unknown third parties outside of the protocol who have zero known incentives or penalties for their performance.
- Chainlink: Natively supports the top leading blockchains through modular external adapters, external initiators, core contract deployments, and token bridging. Nodes directly deliver their data to requesting contracts on any blockchain, minimizing hops, and ensuring timely delivery.
How it Works
Band Protocol v2 (BandChain) is a Tendermint-based blockchain using a BFT (DPoS) consensus algorithm and operating within the Cosmos ecosystem. BandChain nodes are required to operate as both the validators of the blockchain (producing blocks, validating and ordering transactions) and as the oracles that service external data requests. The top 100 validator candidates with the most tokens staked to them become the validators of BandChain. The network’s annual inflation rate will begin at 13.5%, targeting a staking ratio of 66% of the total supply. If the staking ratio is below 66%, inflation will rise to a maximum of 20%. If the staking ratio is higher than 66%, inflation will decrease towards a minimum of 7%. Additionally, two percent of each block reward is diverted to a community fund pool, although it is unclear how exactly these funds will be distributed and to whom.
A validator’s stake can be partially slashed for excessive downtime of more than 30,000 blocks (0.01% penalty), double signing blocks (5% penalty), or unresponsiveness (unclear percentage). If slashed for downtime, nodes will not be able to participate in signing any blocks until they send a transaction on BandChain proving they are online. It’s important to note that BAND staking and all of these slashing conditions are only for maintaining blockchain validation and block production, not for ensuring any oracle service reliability or data accuracy. While the Band team has suggested coinvoting in the future to punish validators performing faulty oracle services (which is subject to whale manipulation), this has yet to be developed and remains entirely theoretical.
Users can send data requests to BandChain by calling available “Oracle Script” smart contracts, which are on-chain executable programs that encode data request details such as the data source APIs to be called and the method of aggregating validator responses into a single response. The nodes chosen to respond to data requests are decided by a randomized stake-weighted algorithm. The more BAND tokens staked by a validator, the more likely they are to be assigned to a data request. Thus, data requesters have no control over which nodes will be used to fulfill their data query.
The chosen validators complete data requests by fetching off-chain data and delivering the response on-chain (to BandChain). The data from each participating node is then aggregated to form a final response. An oracle data proof (Merkle proof) is created and shows the existence of the final result in addition to the oracle script hash, input parameters, time of execution, etc. This proof can be sent to smart contracts on other blockchains to verify it’s attestation by BandChain validators. The proof can also be decoded so the data result can be consumed by smart contracts.
Band Protocol aims to be blockchain agnostic, where oracle requests are processed in BandChain’s environment and bridged on to other chains. They plan to leverage the unreleased Inter Blockchain Communication (IBC) protocol being developed by Cosmos to bridge data to other chains, although it’s unclear if and when IBC will be production-ready. Because of this, users are required to use a lite client protocol to manually bridge data to other blockchains until IBC has launched. This requires users to rely on external third parties outside of the BandChain protocol to actually deliver the data to the consuming smart contract on another blockchain (e.g. a DeFi application on Ethereum). These third parties have no incentive to deliver the data reliably or any punishment for not doing so.
Now that you have some context on how BandChain works, we can dive into an in-depth comparison, going point by point, to see how Band stacks up against Chainlink.
Data Quality and The Range of Oracle Connections
Band Protocol only supports free open APIs. Their support for paid password-protected APIs is at this moment entirely theoretical. The current proposal will require data providers to be individually onboarded, and modify their business model to be paid in cryptocurrency. Chainlink natively supports any and all data sources today, including open and authenticated APIs through modular external adapters that don’t require data providers to change their existing business model.
Band
Band currently only supports free, low-quality APIs, which are commonly updated at 10-minute or more intervals with an accuracy of less than 1 decimal point. The inability to support high-quality data raises serious doubts whether DeFi protocols will feel comfortable relying on Band to secure millions of dollars of user deposited funds, especially when nearly all high-quality data exists behind paywalls and only a single oracle attack due to low-quality data can end up bankrupting and destroying the entire reputation of an application.
Band has stated that BandChain won’t be able to support authenticated (password-protected) APIs until Phase 2+, which has an unclear timeline. The details of how password-protected APIs will be implemented are mostly non-existent, minus a small sentence in their docs stating “API providers to monetize data with BandChain’s on-chain payment settlement”. While it’s unlikely this model is anywhere near production-ready, thinking holistically about it exposes several flaws in logic.
First, data providers will need to change their existing business models to support cryptocurrency payments, meaning Band would have to on-board hundreds to thousands of niche, application-specific authenticated data sources that data requesters require to build their dApps. This is both unrealistic and unscalable, and will severely limit developers from accessing the vast majority of data sources. Second, their randomized stake-weighted node selection mechanism means each node in the network can be potentially selected to service a data request, which results in each node needing access to each potential data source. This would require intense negotiation with the data providers to work out deals where the entire blockchain receives access, as well as slow, complex off-chain coordination/governance between nodes to collectively add new data sources and fund it, making it very impractical or entirely unscalable.
Because of this design choice, Band is only really suitable to support open APIs since it’s the only data type where all nodes can equally access. This will prevent developers from getting access to the specific external data they need, especially in any type of quick frictionless manner. It will also eliminate a vast majority of data providers from ever joining the network as they cannot (for legal reasons) and/or do not want (for business reasons) to accept cryptocurrency payments. Given the lack of applications utilizing BandChain in-production, it will make little sense for any data provider to expend the time, resources, and money to rework their business model to support BandChain. They also have never detailed any features for how to protect data privacy, which will further remove BandChain from being supported by data providers wanting to protect their IP and enterprises needing to protect sensitive information.
Chainlink
Chainlink already leverages credentialed data sources via its Price Reference Contracts — a collection of over 70+ DeFi price feeds secured by decentralized oracle networks that enable any smart contracts to synchronously pull external data within a single transaction.
Chainlink can provide high-quality data in a flexible manner because nodes have always had built-in credential management capabilities that allow them to access any password-protected APIs using modular external adapters. Chainlink external adapters can be written in any programming language and hosted separate from the Chainlink node itself. This opens up the entire world of APIs such as premium data providers, web APIs, enterprise systems, cloud providers, IoT devices, payment systems, other blockchains (both public and private), and much more.
Data providers do not need to change their core business model or backend infrastructure to become accessible to smart contracts because Chainlink nodes can handle API password keys in an external adapter. Thus, they can pay for access to authenticated APIs in the fiat subscription models used today, and then simply place their API password key in the external adapter and immediately start delivering that data on-chain. Data providers prefer this model, especially initially, as they can capture more revenue without expending any time, resources, or money on complex blockchain integrations.
This greatly reduces developer friction because nodes can quickly build support for any off-chain data resource without any permission or needing all other nodes to obtain the same access. All it takes is one node or developer to write an external adapter for that data source API to make it available to any blockchain. Nodes can then differentiate themselves by specializing in offering specific data sets. Similarly, data providers are able to run nodes themselves and sell directly to the market if they wish to get paid directly on-chain. In fact, several data providers are already doing so on Chainlink, such as Huobi, Kaiko, Alpha Vantage, and many more. Importantly, the external adapter model ensures data providers always get paid and enterprise systems only authorize verified users.
Chainlink is also highly aware of the need for oracle privacy, hence their well established work on trusted hardware solutions like Town Crier. Enterprises can use this technology to meet privacy regulations while still using their proprietary data in the execution of smart contracts. On top of that, enterprises can choose exactly which oracle nodes fulfill their data questions, which provides extreme flexibility in ensuring that confidential data (PII), even while encrypted, only stays within a specific geographic region in order to comply with data regulations such as GDPR — a fundamental requirement for many global enterprises and institutions.
The Chainlink network operates as a capitalist free market economy where anyone can spin up their own node to start providing smart contracts on any blockchain with access to any off-chain data source API. This fosters healthy competition and allows nodes to build their own business model without external dependencies on all the other nodes. Conversely, BandChain takes more of a communist, centrally planned economy approach where every node in the network is required to be standardized and support the same features and connections as every other node on the network, ultimately limiting innovation, competition, and adoption.
Reliance on Randomness
BandChain is utilizing randomness as a core element of their oracle mechanism, which is a poorly thought out and dangerous design choice that will severely limit flexibility. It is backed by a team with no experience in mission-critical distributed systems and lacks any supporting academic research. They also got caught trying to copy Chainlink’s Verifiable Randomness Function (VRF) approach to randomness, which was built as a secure input for blockchain games, and backed by world-leading academics.
Band
BandChain uses randomness as a core fundamental aspect of how their network operates, where data requesters are not allowed to choose which nodes fulfill their query. Instead, they are forced into using BandChain’s randomized stake-weighted selection algorithm, which inherently requires that users trust every node in the network.
In addition to the data quality limitations resulting from randomness (mentioned above), their design choice introduces a situation where even a few faulty nodes can generate large security risks for the entire network, of which smart contract creators and data requesters cannot mitigate against. There is no form of pre-emptive quality control or security reviews on node selection, which leaves data requestors fully exposed and puts user funds at great risk of total loss. A node may have a lot of BAND tokens staked to them, but that does not mean they have properly secured their node’s private key (used to sign off on data) or maintain a reliable operation.
Not only does Band use randomness as a core function at the base layer, but the core development team doesn’t even know how to build secure randomness, hence why they got caught trying to copy Chainlink’s solution to on-chain randomness known as Chainlink Verifiable Randomness Function (VRF). They openly stated on their project’s Github repo “Research what and how vrf work in chainlink and how we implement this or our chain”.
What developers or users are going to trust large amounts of value to developers who incorporate core features to their protocol that they do not understand, have backed by no academic research or security audits, and require copying solutions from competitors? It’s yet another example of how BandChain takes existing technologies, blockchain and randomness functions, and tries to force them into being an oracle solution on faulty assumptions.
Chainlink
Chainlink allows data requesters to choose exactly which oracles they want to use, bringing maximum flexibility to smart contract developers. The Chainlink team has built Chainlink VRF to be used as a provably fair source of randomness as an input for the blockchain gaming and NFT industry, not as a core part of the Chainlink protocol itself. They understood and avoided the severe dangers of forcing data requesters to use randomness to choose nodes, specifically the inability to get high quality or permissioned datasets (as described in the data quality section) and massive security risks of not being able to get deterministic guarantees on node operator reliability or infrastructure security.
Chainlink VRF is already being widely adopted, with numerous integrations in a short amount of time after launch, including: PoolTogether, Digitix, Get Protocol, Blocklords, Polyient Games, BlockPegnio, Vibe, Texel, Tiny Boxes, Cargo, Wildcards, FaceGolf, and more. Chainlink’s VRF is built upon years of academic research from industry experts with decades of experience in distributed systems and advanced cryptography, such as Ari Juels, Andrew Miller, Evan Cheng, Alex Coventry, and more (as detailed in the section below). Chainlink VRF was not only developed in-house from some of the smartest minds in the world, but it has been audited by the top security firms and reviewed numerous times before ever being launched into production.
Development and Network Effects
The Band team lacks experience in distributed systems, advanced cryptography, and building oracles, having abandoned their initial v1 launch, which was unable to capture any network effects, and Band v2 being completely unproven with minimal in-production usage. The Chainlink team has been building oracles since 2014 and has a team of over 40+ people, which includes top tier academic and business advisors. Chainlink also secures over $1B+ in value on mainnet for market-leading DeFi projects and has more than 240+ projects (which includes 60+ blockchains) in the integration pipeline.
Band
Band v1 was completely abandoned development-wise and has not seen any in-production usage by external projects. This is evidenced by the lack of any mainnet adoption and the infrequent updates of its price feeds, many of which often showed large deviations. Band Protocol has also changed in design multiple times, including abandoning its dual token bonding curve design, plasma L2, gas fee burn, stablecoin fee support, user-driven data governance, etc. The poor performance of v1 and the constant changes to the protocol puts into question whether the Band team really thought through the design of their oracle mechanism and if they can be trusted to secure potentially millions of dollars of value in other users’ funds.
Band v2 (BandChain) is entirely unproven as an oracle system. Most of the blockchain-related work for BandChain was outsourced to the Cosmos SDK and Tendermint. This lack of time-tested security for any significant amount of real-world value means Band still needs to prove itself as a reliable solution before any projects can feel comfortable trusting its oracle implementation to secure core protocol functions. It’s important to remember that oracles are responsible for the mission-critical functions of other dApps, such as issuing loans and executing liquidations.
Being a nice person or having a degree on your wall is rather meaningless when others are considering whether or not they can entrust the correct functioning of their entire protocol to your external oracle mechanism. Considering Band has minimal amount of in-production users, has changed in design numerous times, and only launched a blockchain using an SDK that abstracted away all the tough work, any dApps looking to use BandChain are taking on serious risk and being borderline reckless with their users’ funds.
The Band team consists of three founders who have no previous experience building blockchain infrastructure, oracle mechanisms, or any mission-critical systems that handle real value. The only experience they have in software development is creating a crypto based mobile game in 2015, which has since been abandoned. Building mobile games for children is not remotely in the same league as building complex distributed systems that not only secure value for Band token holders, but want to eventually secure the value of other applications as well. This experience has very little to no crossover, as building a blockchain oracle network is an exponentially more challenging problem to undertake, with many non-obvious pitfalls and attack vectors. Not only do they lack experience, but they have no academic or business advisors to guide them along the way. All of this points to the Band team being in way over their head, which is going to introduce substantial risks to anyone using it, opening up a possible Mt Gox type implosion when vulnerabilities are inevitably found and exploited.
Chainlink
The Chainlink co-founders have been studying and building externally-connected smart contracts and blockchain oracle technology since at least 2014 when SmartContract LTD was incorporated (before the launch of Ethereum). The team has worked directly with top developers and enterprises who need blockchain oracles, and they understand, through the development of proven solutions used in-production, how to meet the security, privacy, and scalability needs of a broad range of different use cases and applications. They have never abandoned their original vision and instead continue to innovate upon the solid foundation created, including original research developments like TownCrier, Threshold signatures, Mixicles, and more.
Chainlink’s 40+ person development team features seasoned experts in blockchains, oracles, cryptography, machine learning, artificial intelligence, and business development. This includes numerous academic and business advisors which have been crucial to Chainlink’s technology and defense in-depth approach.
Ari Juels
- Professor of Computer Science at the Jacobs Institute at Cornell Tech
- Former chief scientist of RSA
- Formalized Proof of Work consensus in 1999 (powers Bitcoin and Ethereum)
- Invented Proof of Retrievability in 2014 (powers FileCoin and Sia)
- Co-author of the Chainlink whitepaper in 2017 and only works with Chainlink
- Co-author of the Mixicles whitepaper in 2019
- Co-founder of The Initiative For CryptoCurrencies & Contracts (IC3)
- 36,000 total scholarly citations
Tom Gonser
- Founder of DocuSign, the industry-leading e-signature provider in the world
- Joined as a business advisor to Chainlink in early 2019
Andrew Miller
- Decentralized consensus researcher
- Associate Professor at the University of Illinois
- Associate Director of the Initiative for Cryptocurrencies and Contracts (IC3)
- Board member of the Zcash Foundation and Ethereum Enterprise Alliance
- Advisor to both Zcash and Tezos
Evan Cheng
- Former Senior Manager at Apple
- Director of Engineering Blockchain at Facebook
- Co-creator of the LLVM, which generates the low-level machine code running every Apple device, as well as much of Google, Nvidia, and Intel
Hudson Jameson
- Ethereum Core Dev and developer liaison at the Ethereum Foundation (the glue between Eth core devs and the community)
Jake Brukhman
- Former Partner and CTO at Triton Research
- Founder of CoinFund, one of the leading research groups focused on web 3.0 and blockchain-based infrastructure.
Brain Lio
- CEO of Smith+Crown, a widely accepted leader in blockchain research
Chainlink launched on mainnet over a year ago and is already quickly achieving large network effects, currently securing over $1B in value on mainnet from many of the leading DeFi projects including Synthetix, Aave, Loopring, Nexus Mutual, Set Protocol, bZx, Kava, and more. Chainlink also has a pipeline of 240+ integrations in development, ranging from blockchains and crypto startups to global enterprises like Google, Oracle, and SWIFT and enterprise consortiums/standardization groups such as Hyperledger, IC3, EEA, Baseline, InterWork, and more.
The end result is a very experienced and knowledgeable team that is highly capable of solving the difficult challenge of oracles in a manner that is supported by crypto startups, global enterprises, and leading academics. Developers have case studies and definitive proof that Chainlink is a secure and reliable oracle solution for securing high-value contracts through its real-world usage in production. Its adoption is already driving a positive feedback loop where the more users, data providers, and nodes using the network, the more powerful and valuable Chainlink as an oracle solution becomes to new and existing stakeholders. Band has been unable to achieve any network effects given its lack of mainnet adoption and a very limited number of partners (most of which are validators or exchange listings). Additionally, developers have no information about Band’s performance as an oracle besides its lackluster v1 launch where feeds didn’t update for sometimes weeks to months at a time.
Resource Efficiency and Network Design
BandChain uses its own blockchain, which is completely unnecessary to the job of an oracle — delivering data to smart contracts. The Band team is using blockchain solutions to solve a fundamentally different problem, which leads to numerous issues such as state bloat, long sync times, and forced random node selection. Chainlink is not and will never be its own proprietary blockchain, but instead is truly blockchain agnostic middleware, which allows it to be flexible to support all DLT networks and data resources, as well as remain extremely lightweight for node operators.
Band
The design choices behind the BandChain blockchain (namely the two second blocktime) results in a quickly growing state bloat for all nodes and a longer sync time for new or reconnecting nodes. Running a Band node is much more expensive and resource intensive than blockchains such as Bitcoin and Ethereum, which have been optimized for lowered hardware requirements in order to increase decentralization. It is also exponentially more expensive and resource intensive when compared to Chainlink oracle nodes, which do not operate a proprietary blockchain at all.
It’s more than likely that BandChain’s design decisions will naturally lead to increased centralization over time, with nodes mostly run in a few data centers by well-capitalized validators. These are the same issues that blockchains like EOS and BSV have been affected by. In order to cover the cost of running a node, BandChain will need a higher inflation rate of newly minted tokens to maintain the same level of security over time if user fees do not increase to a high enough level.
Since Band’s oracles are also block producers, a failure in block production (like the recent memory leakage and state pruning issue in the Cosmos SDK) leads to a failure of the oracle mechanism too. Integrating an oracle mechanism directly in the base layer of a blockchain also forces a single monolithic network design, which hinders oracle innovation due to enforced standardization and greatly limits the oracle’s generalization in being able to support a wide variety of use cases (as showcased in the data quality section).
BandChain is effectively operating through a security through obscurity approach, where validators have no requirement to reveal their identity or reveal their reliability. This combined with the randomized stake-weighted selection means data requesters have to trust all nodes operating on BandChain as they have no control over which anonymous or non-anonymous nodes will end up getting selected, creating a high trust barrier for users. This is far from ideal as security through obscurity is generally accepted by most security experts to be a poor way to design a fundamentally robust system or network.
Chainlink
Since Chainlink is not a blockchain, it does not suffer from state bloat or sync time issues. Chainlink nodes are lightweight with extremely low hardware requirements, which are able to operate in any compute environment through a simple Docker container. Nodes have even been set up to run on extremely low spec devices such as a Raspberry Pi, making it so IoT devices can run a Chainlink node directly to reduce the number of hops to a smart contract. This low barrier to entry from a lack of resource requirements greatly increases the decentralization of the Chainlink network overall.
Unlike BandChain, there is no need for monolithic global consensus between all active Chainlink nodes, instead opting for horizontal scalability where an unlimited number of oracle networks run simultaneously in parallel. Chainlink nodes do not need to create or validate the blocks of any proprietary blockchain, thus there is no dependency on the functional operation of any one specific blockchain. Chainlink is completely self-sovereign in its security for each individual oracle network; it does not make any implementation assumptions and instead offers the tools needed to create any type of oracle network using whichever nodes, data sources, or parameters are desired for any blockchain. It is highly generalized, meaning it does not impose any one type of oracle design pattern on users (push or pull, permissioned or permissionless, centralized or decentralized, open or whitelisted, deviation threshold and/or heartbeat updates, etc), allowing for limitless innovation on new oracle models.
Chainlink networks currently running on mainnet operate using a security through transparency philosophy, where the Price Reference Data networks consist of only security reviewed nodes which are operated by experienced blockchain DevOps and security teams distributed around the world in data centers and on-site bare-metal servers. Additionally, multiple analytic services have been created by numerous independent teams that allow end-users and developers alike to peer directly into the performance and history of each oracle network. This enables anyone to definitively know whether a specific Chainlink network or node is reliable or not using historical indexed on-chain data presented in a simple, easy to consume fashion. This includes feeds.chain.link (created by Chainlink), market.link (created by LinkPool), reputation.link (created by Secure Data Links), and more.
Decentralization and Crypto-Economic Security
BandChain uses a single monolithic consensus mechanism with an upper bound of 100 nodes due to the limited scalability of Tendermint BFT. Validators only stake to produce blocks and are not slashed for manipulated or invalid external data. Chainlink is a collection of independent decentralized oracle networks operating in parallel across numerous blockchains, with the threshold signatures and off-chain aggregation upgrades having no upper bound on the number of nodes in each network. LINK staking will involve binding service agreements with pre-defined slashing conditions on data quality, reliability, accuracy, etc.
Band
Since BandChain operates using Tendermint BFT consensus, it can only scale to around a couple hundred validators at most before communication bandwidth becomes an issue and the Transactions Per Second (TPS) tanks. For this reason, Band has a limited validator count up to a maximum of 100 nodes at any given point in time. This is another downside of using a monolithic blockchain as the sole consensus mechanism for processing data requests and responses. It greatly limits the maximum level of decentralization that BandChain will ever be able to reach, barring a fundamental redesign of the entire network model.
It’s also important to note that BandChain validators only get their BAND stake slashed if they act as a poor block producer (downtime, double signing, unresponsiveness). Acting as a poor or malicious oracle (manipulated data, oracle network attacks, collusion) does not cause any stake slashing penalties. Validators within BandChain are likely to be more focused on being a good block producer rather than being a good oracle because that’s what determines if they get slashed or not.
While retail holders have discussed BandChain staking, the reality is that retail staking to validators doesn’t provide much of any real value to BandChain as an oracle network. In the future, BandChain may implement slashing for nodes delivering bad data through on-chain voting, but there have been very few details as the only mention is a small suggestion as a single sentence in their deprecated D3N spec. Info on the subject has since been removed from the current documentation on their github wiki, showing they likely have little to no active development on the topic.
Although this idea may have also been abandoned (like many other previously core features), they have stated in the past that retail BAND token stakers would help curate data providers as a reputation system mechanism. This represents a very weak and borderline dangerous form of enforcing quality control, especially when considering it's supposed to provide security for contracts worth millions of dollars or more. Retail investors are unlikely to perform any serious due diligence and will just choose the data providers with the best immediate ROI.
Chainlink
The Chainlink team has been actively developing Threshold Signatures to enable off-chain aggregation and batching of node responses within an oracle network into a single transaction to enable cost-efficient decentralization. With threshold signatures, Chainlink oracle networks can scale to an unbounded number of nodes, enabling extremely high levels of decentralization with extremely minimal gas costs. In combination with the FluxAggregator and FluxMonitor, network congestion on any blockchain becomes a non-issue for any sized Chainlink network.
Not all staking based protocols are created equal, especially for oracles. Chainlink’s staking mechanism is designed to be much more robust, generalized, and focused on oracle security/integrity when compared to Band’s retail validator staking. Chainlink is implementing support for binding Service Agreements, which are smart contracts between the requesting application and the node operator(s) that are cryptographically signed by both parties. Service Agreements outline the exact terms of service that the node operators in a network need to provide and the penalties for not upholding those terms.
The service agreement is backed by staking from node operators, who deposit a predefined amount of LINK tokens as collateral to back their oracle services (how much is negotiable between the node and data requestor). The collateral is only released if they fulfill their obligations. Conversely, if a node does not fulfill its agreed-to obligations, its collateral can be slashed according to the predefined terms in the Service Agreement. Slashing conditions can range from non-responsiveness, providing manipulated/outlier data, and even customized conditions or multiple different levels of slashing for varying levels of punishments. This gives data requesters an extreme level of flexibility in how edge cases are handled and how nodes will be punished financially.
To put it simply, Chainlink’s staking will ensure high quality and accurate oracles, while Band’s staking only ensures honest block producers. BandChain was built as a blockchain first and an oracle mechanism second, while Chainlink was built from the ground-up to be a generalized and highly flexible oracle framework that can be used to bidirectionally deliver data to any blockchain.
Approach to Blockchain Agnosticism
BandChain relies on the yet to be released Cosmos IBC to natively bridge data to other chains. It currently only supports lite clients where data is required to be delivered by unpaid and potentially unknown external parties outside of the protocol. Chainlink natively supports the top leading blockchain today and will continue to support future chains through its adaptive and easily integratable design.
Band
Since BandChain is its own blockchain, any decentralized application on another blockchain that wants to use Band oracles is completely reliant on a cross-chain communication bridge. BandChain’s primary cross-chain functionality is dependent on the Cosmos Inter Blockchain Communication (IBC) protocol, which isn’t being built by their team and is still months to potentially years away from being fully operational and time-tested, especially for trustless data transfers from one blockchain environment to another.
Until then, blockchain interoperability will be performed through BandChain lite clients deployed on each integrated blockchain. Smart contracts on other blockchains using Band as their oracle will be dependent entirely on some external non-incentivized third party delivering proofs to the lite client contract on that chain. BandChain validators do not deliver these proofs to other chains (not until IBC), meaning a third party is currently always required to relay the data between chains. Such an approach is difficult to scale as each supported blockchain will need to host its own lite client contract and have external entities outside of the Band protocol delivering the proof and keeping the lite client synced.
There are major questions about the incentives for delivering the data to lite clients, especially during high network congestion when transaction fees skyrocket on the external blockchain. It’s important to note that even though BandChain aggregates data on its own blockchain, broadcasting the data itself to other blockchains will be subject to the consuming chain’s native speeds and latency. Bandchain currently only supports a select few testnet blockchains.
Chainlink
Chainlink is blockchain agnostic, but achieves such functionality in a fundamentally different way than BandChain. Instead of operating its own proprietary blockchain, Chainlink is a framework for building heterogeneous decentralized oracle networks for smart contracts on any blockchain. Through modular external adapters and external initiators, Chainlink nodes are able to monitor data requests and deliver off-chain data to any blockchain environment directly, meaning there is no reliance on lite clients or external third parties with questionable incentives/accountability. Thus, Chainlink nodes are able to natively support any blockchain, DLT, or layer2 solution that exists today or into the future, including chains that have limited smart contract capabilities like Bitcoin where such a lite client could not be hosted. This also encompasses non-IBC compatible permissioned chains and non-blockchain systems like enterprise ERP/CRM backends and legacy databases that want to create and consume external data from a crypto-economically secured, decentralized oracle network.
Supported blockchains on the Chainlink network include Ethereum, Bitcoin, Polkadot, Ava, NEAR, Kava, Tezos, Binance Smart Chain, China’s BSN, IRIS, EVM based chains, Substrate based chains, Harmony, Zilliqa, Ontology, Kadena, Klaytn, CasperLabs, Solana, Hdac, ICON, Conflux, Hedera Hashgraph, Matic, and many more. There are over 70 blockchains which are being integrated into the Chainlink protocol. While initially launched on the Ethereum blockchain, Chainlink’s systems of contracts are being rewritten to run natively on numerous other blockchains with the LINK token being wrapped via a cross-chain bridge. As a result, each Chainlink oracle network will only be subject to the throughput and security assumptions of the particular chain it’s running on.
Chainlink’s form of blockchain agnosticism can scale much higher than BandChain, even with IBC support (of which Chainlink can also support). Instead of Bandchain’s approach of funneling data requests and node responses through a single blockchain, the Chainlink network can process data requests and node responses in parallel and completely independent of one another, offering theoretically infinite horizontal scalability. Chainlink is blockchain agnostic in every sense of the word, due to the fact it is not a blockchain, but an abstraction layer running simultaneously on all blockchain networks.
Conclusion
Chainlink measures twice and cuts once, practicing agile development with the most renowned researchers in the blockchain industry, in order to stay on top of new cutting edge technologies and innovations. The team recognizes the importance of a well thought out and end-to-end secure oracle framework that mitigates complex attack vectors ahead of time. Band’s approach is more akin to throwing spaghetti at the wall and seeing what sticks, or in other words, following a ‘move fast and break things’ philosophy. While this approach can work for some industries, it is absolutely not a desired approach to creating a solution for the oracle problem where absolute security is a must and real value is vulnerable to manipulation that cannot be reversed.
When fully compared, Band Protocol is going for a blockchain-based approach to solving the oracle problem by using monolithic consensus, which enforces standardization of nodes and has upper limits to its decentralization. This rigid model is why Band will struggle to support authenticated APIs in a scalable manner, as it’s near impossible for all nodes to support credentialed data in their framework without redesigning the whole data industry from the ground up. Most of their original development and design spec has been abandoned and future feature sets are almost entirely theoretical, extremely vague, or unproven to secure real economic value. While exploration and changes can be good at times, they can also be a sign of someone that is in way over their head, which should be especially concerning with what’s at stake when building an oracle protocol.
Alternatively, Chainlink has determined that the oracle problem is separate from the blockchain problem, going with a highly modular, free-market framework supporting an infinite number of parallel oracle networks that reach consensus and maintain security completely independent of one another. Chainlink users have access to high levels of customization on multiple fronts including node selection, data types, security approaches, collateral staked, privacy support, and more. The Chainlink team has proven its ability to support a large growing ecosystem of DeFi projects on mainnet, securing over $4B+ in user funds. Chainlink is already building public good oracle networks where multiple users contribute to the same price feed lowering the costs for each user.
While competition is healthy and everyone should do their own due diligence, it’s important to really examine the oracle mechanism responsible for directly securing large amounts of user funds. Oracles are not something to take lightly if the smart contract economy expects to scale and be utilized by traditional institutions with much more conservative risk appetites and higher demand for flexibility and quality control. A ‘good enough’ approach to oracles is exactly the mindset of developers who get hacked and lose their users’ funds. It’s no different than an unaudited smart contract being deployed, which is exposed to massive unknown security risks that could completely bankrupt a project and its users. Similarly, a non-generalized solution may serve a single niche in the short term, but will be swallowed up by a generalized standard as is the case in many tech industries.
It’s for these reasons of needing to scale security up with smart contracts’ value and generalization powering superior network effects that we believe Chainlink is by and far a vastly superior oracle solution when compared to Band Protocol.
Follow us on Twitter @SmartContent777 to get up to date on the latest articles, as well as follow our individual accounts @Crypto___Oracle and @ChainLinkGod for a constant stream of information about the Chainlink, DeFi, and the blockchain space.
