As the core role of whole system, Keeper Network is responsible for the custody of underlying crypto assets and provides crypto collaterals against system security risk.
Running nodes is another responsibility for Keeper. They need a cloud server to configure private key of Ethereum Network to verify identity and interact with smart contracts, and to generate a pair of keys to initiate and verify Bitcoin multi-sig transactions.
Smart contracts deployed on target blockchains will mint the relevant pegged value crypto assets. The whole process is organized in a 100% decentralized way given Keepers are permission-less to join or leave the Network at will. Users can cross chain their crypto assets in a secure and decentralized way up to a capacity backed by the crypto collaterals provided by the Keepers Network.
Due to the exclusive algorithm of DeCus, Keepers are managed to be decentralized custodians at a low collateralization rate, with the security of custody still being guaranteed, thus improving the utilization rate of crypto assets.
DeCus is launching with tokenized bitcoin-eBTC. The following picture shows how bitcoin can be turned to eBTC.
One user who wants to mint eBTC should first submit a request to the smart contract on Ethereum network and then get a multi-sig BTC address of a designated keepers group. After depositing a certain amount BTC to that multi-sig address and submitting the proof of deposit, the corresponding amount of eBTC will be sent to the user’s ETH address.
Users are able to convert their eBTC back to BTC at any time. The redeeming process is similar with minting. Once the original BTC is received, the corresponding eBTC will be burnt.
A decentralized and self-regulated Keeper Network based on Dr. Yang Guang’s cryptography IP is the cornerstone of whole system.
Suppose the BTC assets are kept by 10,000 custodians in a multi-sig address from which every transaction needs at least 5,001 signatures. As long as the majority of custodians remain honest, an adversary cannot steal any BTC under custody. In this case, a minimum collateral would be sufficient to enforce correct behavior of custodians.
However, a 5,001 of 10,000 multi-sig address would be technically challenged to use on Bitcoin network, and the composition of a transaction with 5,001 signatures would require non-trivial coordination among all custodians.
To balance the security provided by large groups and accessibility by small groups, DeCus introduced a custody scheme based on overlapping group assignments. The following toy example will show how it works.
Suppose there are six custodians partitioned in two 2-of-3 multi-sig groups each of three custodians. Then by corrupting two custodians the adversary is able to get full control of one group, which is 1/2 of the total assets under custody.
Alternatively, let’s enumerate all 2-of-3 multi-sig groups drawn from the six custodians. There are 20 groups in total, whereas by corrupting two custodians the adversary only controls four groups (each corrupted group consists of two corrupted custodian and another honest one). Thus, the same 1/3 fraction of adversary only controls 4/20=1/5 of all groups. In other words, the new scheme is able to securely keep 5 BTC with total collateral of 3 BTC against 1/3 adversary, since the profit of launching an attack does not exceed the adversary’s collateral.
Of course, the above toy example is designed for illustration but not impressive collateral efficiency. We can estimate the ratio of collateral to total assets under custody asymptotically as follows:
Based on the above analysis, DeCus implements the decentralized custody scheme with fewer groups in representative, and achieves impressive low collateral ratio.For example, with a polynomial design over prime fields, DeCus is able to achieve efficiency factor η=A/C as much as 20 with n=121 custodians, which means the collateral is <5% of total assets under custody!
Keepers are responsible for the custody of underlying crypto assets. All keepers are organied under the overlapping group assignments in a 100% decentralized way , and they need to deposit collaterals in case of misbehaviors.
DeCus will soon launch a permission-less Dutch Auction to select the first batch of Keepers. Anyone who is interested to be a Keeper can participate the auction without any KYC or AML processes.
After several rounds of iterations and the system enters the mature operation stage, Keepers can join and leave freely under the new mechanism. For every user's deposit, a new Keepers group is pulled together (selected by the random beacon), and they generate a Bitcoin multi-sig address for the depositor, which is marked on the Ethereum chain.
The stable operation of system relays on Keeper Network. If the number of Keepers is too small, the whole system will become insecure. So we need an efficient and fair way to initiate a bootstrap process of the Keeper Network. Under current circumstances, Ductch Auction seems to be the best way because of two features: 1) Permission-less, which means anyone who wants to be a Keeper can participate in the auction; 2) Transparent, which means all participants can compete fairly under the set rules; 3) Participation discovery, which means the real willingness to become a Keeper can be figured out through the whole process.
During the auction, candidates need to set the amount of deposit they would like to commit, according to which the system will rank in real time. By the end of the auction, the top n will be selected, with the n-th committed deposit amount being used as the final deposit amount for each selected keeper. During the auction stage, the deposit can be retrieved at any time. Once auction ended, the mint stage will launch at which the Keepers' collateral will be migrated to the Vault of Keeper Network and each Keeper operates his/her node of the Keeper Network.
As the core role and long-term partners of whole system，Keepers will receive mostly 3 parts of benefits: 1) Share of system revenue, which is generated from the fees of eBTC(or other pegged value crypto assets) minting and redemption. 2) Collateral interests, namely the deposit interests of Keeper collaterals. eBTC is the first tokenized bitcoin that gives such interests to the custody party of underlying assets. 3) Farming yields, which will be generated by integrating crypto collaterals into yield farming protocols such as Curve, Compound and AAVE etc. Whether authorize the system to farm is determined by Keeper itself.
Decentralization and scalability are well balanced by eBTC compared to other tokenized bitcoins.
Some tokenized bitcoins, like WBTC and HBTC, are issued by a centralized party, so they may become irredeemable in case of a single point failure or censorship.
Other BTC-backed tokens, such as tBTC and renBTC, are issued in a decentralized way. However, they rely on a high collateralization rate to mitigate the risk of price fluctuation and misbehavior. For example, each minted tBTC-token requires collateral of value ≥1.5BTC, and each renBTC requires collateral value ≥3BTC. Such over-collateralization would significantly increase the cost of using tokenized bitcoins.
eBTC gives a different answer. The decentralized and self-regulated Keeper Network of the system enables eBTC implementation at a much lower collateralization rate, with 30%-50% compared to the average more than 150% of other solutions, while still kept decentralized.
Given Decus as a decentralized custody system and eBTC as an underlying asset protocol within the Defi building blocks , it's intended that DeCus starts as a community-driven initiative from day one, and eventually goes to DAO governance.
Community members can propose ideas and votes on proposals in DecusDAO where DeCus shall head towards. Once the proposals are passed, DeCus team will be responsible to carry them out.
Of course, all of our team members are open and willing to collaborate with anyone wanting to build with together. It would be amazing to see leading developers and DeFi protocols participate in this program.