ROTOKEN

Author

Aleksandar Veljković

Published

March 3, 2022

Executive Summary

Organisations may include non-founder members with specific roles (treasurer, secretary, board members…) with higher permission levels, compared to the regular members. Members that have permission levels which include control over organisation finances represent a high risk for the organisation as there is a high risk of losing money if malicious members perform illegal transfers. That problem is especially present in the world of cryptocurrency where transfers cannot be revoked, once transaction is executed there is no rollback. The use of multisig might be an obvious solution, but the high overhead cost of writting multiple signatures on blockchain, even if the transaction is valid, may not be acceptable in many situations where there is a high frequency of transactions.

This research proposes a solution for this problem, restricted to the world of crypto-financial transactions, by introducing a rotating token which guarantees randomised selection of members that would be granted specified roles over fixed period of time while allowing blockchain money-transferring transactions to be executed only by the members that are holding the token. For improved prevention of unwanted transactions, the transactions may be disputed before their execution, which allows higher control over money transfers and enables improved protection of organisation funds. All transactions are transparent to all approved members, which adds extra layer of transparency over the entire organisation budget.

Introduction

Money transfer transactions are the essential parts of any organisation. Protection of organisation finances and prevention of malicious transactions is one of the top priorities. There is usually one, or a few people that are responsible for all money transfers in the organisation and are often the only members that are actually aware of those transfers. Malicious member with access to organisation funds may use the opportunity to leak money from the organisation for the causes that are against the organisation rules or even against the law. Board members of the organisation are ofter not aware of those malicious activities as the members that perform those transaction also prepare financial reports, hiding their tracks. An example would be a fake invoice representing payment for the fictive service that actually represents money laundering. Even invoice payments for the regular services may be modified with highly increased price.

Current mechanisms for controlling financial transactions in smaller organisations basically do no exist as every lost cent is easily trackable when dealing with smaller amounts of money. When dealing with large number of transactions and bigger funds, the situation is not that clear. Bigger organisations specify budgets for certain topics, restricting the maximum amount spent on specific topic, therefore defining the upper limit on transaction fraud potential [1]. Those budgets are not small, Microsoft Corporation spent more than 20 billion USD on sales and marketing in 2021 [2]. There is a big potential for hiding malicious transactions.

Choosing members that would be granted access to funds is currently solved in two ways: - Hiring accounting department [3]. In this setup, the employees are hired for the longer period of time, sharing the responsibility with each other and the head of the department. - Electing treasurer [4]. In this setup, the treasurer role can be assigned to different members, often by voting. The job of the treasurer is a person responsible for observing the accountant department, but in smaller organisation he is often the only person dedicated to manage and observe financial flows. The issue appears when the malicious clique tampers the voting and forces the election of a certain trusted member as long as they all benefit from it. There is also the case where the treasurer role is assigned once and never changed, assuming his honest work.

To prevent the problem, it is important to understand its cause. Newly granted members with access to organisation funds are often working by the book, scared of possible consequences when caught performing a malicious activity. As the time goes by, their actions are not carefully observed and their position in the organisation becomes stable. At that point, the risk of malicious transactions becomes higher as the actors are not observed and there was enough time to form a clique with other members [5,6]. To quote Margaret Murray: “Governments are like underwear. They start smelling pretty bad if you don’t change them once in a while.” The same goes with any position of power, even if that power includes just access to finances.

It seems that there are two main problems, unbiased change of authorisation and transparent transaction control.

Crypto-world introduced multisig wallets for improved safety. The problem with this approach is that it introduces costs to multiple members for executing one transaction even if that transaction is not malicious. The better approach would be a procedure that would add extra cost only to prevent malicious transactions, thus saving money when the transactions are not malicious.

Unbiased selection of participants to perform certain actions is demonstrated in certain Blockchains where validator nodes are selected in Round robin manner. Round robin approach may not be adequate for selecting treasurers as it is predictable and assumes that next treasurer will be available to accept the role. In general, unbiased selection should be incorporated wherever there is a high risk involved with knowing the choice in advance or where all choices should be equaly probable. Such cases include distributing role, data or duties to randomly selected entities, selection of randomly generated teams (such as jury selection) and similar.

Goals & Methodology

The main goal of this research was to find a solution for having unbiased selection of personnel with access to finances as well as transaction control. The way to get to the solution was to present first all the requirements and restrictions.

The transactions that are focus of this research are blockchain transactions. This research will focus on Ethereum money-transferring transactions so the base element of the solutions would be Ethereum smart contract, for this research named ROTOKEN. ROTOKEN smart contract includes NFT token and wallet. The wallet is controlled only by the owner of the NFT.

All listed members of the organisation should be automatically whitelisted, meaning that all members should be able to get selected for special roles involving access to finances. This is naturally enforced as the salary payments are also transactions from organisation funds and therefore controlled by the ROTOKEN smart contract.

Unbiased selection can only be performed using unbiased probability distribution over the members that are applying for the position. The selection needs to take into account that next selected member can’t be the same as the previous one. Random selection on blockchain is a tricky part as it relies on values that can be predicted in advance. One of the sources of pseudo-randomness is the block hash, as it is a result of SHA3 function. However, the value is predictable by the miners and it can be tampered so it should not be used as a source of randomness. Another approach is using verifiable random functions such as Chainlink VRF. This might be costly but it is a good solution. This research also proposes another, possibly cheaper approach. Selecting the latest transaction hashes of \(k\) wallets outside of the organisation ,with the most frequent transactions, and hashing them all together may be a good source of randomness as the main problem with predictable randomness from block hash is the ability of the miner to reorder transactions until the desired hash is achieved. By selecting transaction hashes instead of blocks, the only way to tamper with the randomness is to selectively mine transactions which leaves much less room for tampered combinations and transaction hashes are close to random as they are influenced by many changing variables.

Selection of new members with special roles should happen regularly in fixed intervals. As transaction of changing the role needs to be executed on blockchain, it is required that transaction is explicitly invoked. To incentivise regular changes, organisation funds would be blocked for transactions until the role is changed. Role change should not be allowed during before the end of current member’s term unless it is invoked by the current member. The role transfer to a new member is done by transferring ROTOKEN NFT.

Transaction that is submitted on blockchain can’t be revoked. The only way to prevent transaction execution is to prevent its submission. Proposed solution includes two step execution. The first step is transaction announcement, which should be executed on smart contract and contain details about target wallet and the amount being sent. The second step is transaction execution, which submits pending transaction to blockchain for execution. Between two steps, certain period of time should be left for disputes. If the dispute is initiated by enough participants the transaction will be canceled, where enough should be >51% or >2/3, but the exact limit would be configurable value.

To disincentivise submission of malicious transactions, that would eventually probably be disputed but also add extra cost for disputing, submitter should be penalised. For a member to loose money it is required that the member has the money in first place, so the safest way to ensure successful penalisation is to include staking. The money staked by the submitter would be lost in case of successful dispute.

Lost staked money needs to change ownership. If the money is sent to disputers, there would exist an incentive for members to dispute every transaction and earn money. To prevent this behaviour, the proposed solutions is transferring money to the organisation wallet. This solution may seem problematic and unnatural but there is an explanation for that. One way to look at staking is to take existing money out of the pocket. The other way is to take money “from the future”, meaning that member may have dedicated annual fond which is used to issue monthly payments as salary and may also be perceived as staked money. In this case, the penalisation would assume reduction of the future salary which technically is returning money back to the organisation. To further disincentivise disputes from the side of the organisation executives, disputed transaction, defined by the receiver’s address and the amount of money to be transferred, blocks further transfers to receiver’s address for certain amount of time. If the disputed transaction was indeed malicious, this mechanism will block further transactions to that address, preventing new attempts. If the transaction was falsely disputed and the receiver is a proper receiver of organisation’s money, by preventing the transaction and blocking the transfer to the proper receiver, the behaviour directly disrupts regular functioning of the organisation.

Results & Discussion

ROTOKEN smart contract should follow modifier ERC-721 smart contract. Token transfer should be possible only by internal call and the token should be transferred to randomly selected whitelisted member. Additionally, the smart contract should include: - list of whitelisted members, with getters and setters - list of transactions with statuses - dispute function - transaction announcement and execution functions - ownership change function, which would internally call transfer function

Proposed ROTOKEN smart contract interface:

interface ROTOKEN {
    // + ...ERC-721 interface

    Enum { PENDING, DISPUTED, CANCELED, FAILED, EXECUTED } transactionStatus
    struct (
        transactionStatus status,
        address currentTokenOwner
        address receiver,
        uint256 amount,
    ) transactionData;

    mapping(transactionId => transactionData) public transactions
    address[] whitelistedMembers;

    event TransactionStatusChanged(uint256 indexed transactionId, transactionStatus status)

    function announceTransaction(address receiver, uint256 amount) public returns(uint256 transactionId);
    function executeTransaction(uint256 transactionId) public;
    function dispute(uint256 transactionId) public;
    function changeOwnership();
}

Potential markets

Potential markets that might be interested to use ROTOKEN include: - DAOs - Companies with frequent audits by the investors, board members or shareholders, mostly startups - Banks - Government organisations related with public government funds

Conclusion

ROTOKEN smart contract proposes solutions for unbiased selection of members for granting control over organisation funds and incentive based transaction control. Minimal additional operating costs required for essential control are added, on top of regular transaction costs. The smart contact extends ERC-721 interface with certain modifications and extensions which makes ROTOKEN NFT compatible for integration with the existing systems where the ERC-721 tokens are used. Those systems may use the information about the ownership of the ROTOKEN to grant the owners additional privileges, like document signing, voting etc. Next steps include market evaluation, PoC implementation and detailed operating cost estimation.

Bibliography

[1] https://corporatefinanceinstitute.com/resources/knowledge/finance/financial-controls/

[2] https://www.statista.com/statistics/506534/microsoft-sales-marketing-expenditure/

[3] https://www.investopedia.com/terms/a/accountant.asp

[4] https://nonprofitlawblog.com/treasurer-duties/

[5] https://corporatefinanceinstitute.com/resources/knowledge/other/fraud-red-flags/

[6] https://www.nsu.edu/About/Administrative-Offices-Services/Internal-Audit/Making-a-Report/Red-Flags-of-Fraud.aspx