The South African Reserve Bank has released the results of Project Khokha, a proof-of-concept settlement platform based on distributed ledger technology (DLT).
With ConsenSys as its technical partner, the goal of the project was to build a system that handles real-time gross settlements.
Currently, the SARB runs a system called SAMOS (South African Multiple Option Settlement system) to perform real-time gross settlements between local banks.
Project Khokha aims to replicate some of the functionality of SAMOS in a distributed ledger.
ConsenSys provided its Ethereum-based permissioned blockchain product, called Quorum, for the project.
It then connected eight banks to its network, and set up test cases to see how the distributed ledger processed transactions.
The banks were the SARB, Absa, Capitec, Discovery, Investec, FirstRand, Nedbank, and Standard Bank.
Another aim of the test was to gain a practical understanding of the technology.
No real transactions were processed, although the SARB said it made tests as realistic as possible.
Project Khokha tested whether the following is true of an Ethereum-based real-time gross settlement system:
- Real-time gross settlement transactions can be executed using standard payment message formats (ISO 20022).
- Messages can be processed at sufficient scale, and in line with the current system processing times.
- Blocks should be propagated within 1s to a 95% confidence level, and within 2s to a 99% confidence level.
- The confidentiality and privacy of transactions between commercial banks is maintained.
- The visibility of the system for the SARB is sufficient for oversight and operational management.
All of the these requirements were met.
Project Khokha was also able to process the typical daily transaction volume of South African financial institutions within two hours.
The network was configured in such a way that the SARB does not need to check and approve all the transactions.
Instead, the banks approve transactions on the network, while keeping the details of transactions that do not involve them confidential.
This means that even if the SARB’s node is down, banks can continue to make settlements.
Tests also included invalid transactions, where incorrect identifiers were provided, or where the bank did not have sufficient funds to make a transfer.
The SARB said the full results of the work it has done will be released through the course of the year.
“The business case for implementing DLT systems is usually not made on a like-for-like replacement basis,” it said.
SAMOS is no different, and doing a straight replacement with a DLT system equivalent would probably not be clear cut.
“What is more interesting is the broader impact of a DLT system and the kinds of things it enables as applications and use cases are built around it.”
The next step from Project Khokha may be an economic impact analysis of the implementation of DLT under a number of scenarios, it said.
“Such a study would need to consider varying degrees of DLT usage in the economy and would be an important parallel piece of work to further experiments with the technology.”