Building a Blockchain POC for a major financial institution: the capabilities and limitations of existing technologies
1. BRIEF
โข ``To design and build a distributed ledger POC system
to process and store proprietary messages for inter-
subsidiary forex transactions internal to a major Fortune
500 ๏ฌnancial institutionโโ
โข Why? Intermediate message processors add time and
money. Cynically: get in bed with the technology that
may one day supplant them.
โข Requirements? Throughput and decentralisation.
www.kwori.co.uk www.kingsoftheblock.com
2. TECHNOLOGY
โข Con๏ฌicting metrics: transaction throughput vs.
decentralisation
โข โ BigChainDB vs. ErisDB (Permissioned chains
based on Ethereum / EVM) + Tendermint (PoS;
faster)
โข ErisDB brings smart contract technology in the form
of Solidity.
โข Both would be employed along with Meteor (novel
client-server communication) and other popular
frameworks such as NodeJS.
โข Client selected ErisDB.
3. PROBLEMS
โข End product? 4 weeks + N system revisions + 1 proprietary load
balancer = 20 Tx/s. Still several known issues based on Eris bugs.
โข Solidity/EVM memory provisioning
โข EVM stack depth limits function implementation.
โข Solidity/EVM memory management
โข Memory provisioning not suf๏ฌcient for basic data structure operations
(e.g. adding items to an array).
โข Poor documentation and error messages exacerbate everything
โข Result = no smart contracts
โข Instead, multiple blockchains, application-based ๏ฌltering and direct
transactions.
โข Question the role of smart contracts overall in data processing
applications (i.e. processing on the blockchain), at this stage.