When a Blockchain Isn’t a Blockchain
The consulting company Accenture has come up with a way to render the blockchain -- the technology behind Bitcoin -- more palatable to the world of high finance. The innovation: Make it not a blockchain.
One of the defining features of the blockchain is its immutability. Numerous computers retain identical copies of a permanent, encrypted record of each transaction. Nobody can go back and make changes, so no central authority -- neither a government nor a bank -- is needed to guarantee the transactions' legitimacy.
This permanence, though, can be a problem in financial markets. Humans are prone to error, and it would be terrible if such errors could never be amended or undone. Imagine what would have happened in the wake of the 2010 Flash Crash if exchanges hadn’t been able to cancel thousands of erroneous trades.
Enter Richard Lumb, Accenture’s chief executive for financial services, with a prototype for a fully redactable blockchain. It's a modification of the Bitcoin protocol 1 , augmented with a trapdoor function that creates a master key to change or delete any existing entry.
The modification isn't easy: Remember, the record is maintained on numerous computers. Redactability requires the creation of a governing authority, which can make changes and distribute them to all the participants. In the case of Bitcoin, downloading an entire blockchain can take days. Accenture says it addresses the issue by allowing the trusted authority to edit individual pieces of the chain, subject to validation by the participants.
In other words, Accenture has solved the blockchain's immutability problem by creating a giant, horribly inefficient Excel spreadsheet. Which raises another question: When will financial institutions stop to ask themselves why they need a blockchain at all?
Investment banks have high hopes for the blockchain's potential to cut costs in clearing and settlement, which involves making sure money and securities get to their new owners after a trade has been made. But a decentralized ledger -- again, one of the blockchain's key features -- is always more computationally expensive than a central one. There needs to be a very good reason to have multiple computers replicating the work of one.
Erroneous trades, for their part, are pretty easily dealt with in the current system. Most exchanges have procedures in place for canceling inadvertent trades. They can do this because they operate in a regulated industry where a central authority ensures the legitimacy of transactions. A blockchain is a place to create protected transactions when they can’t be protected in any other way.
Most of the decision makers in financial institutions don’t understand computer systems, making them easy targets for people peddling blockchains. This is aided by the fact that banks themselves continuously advertise their own blockchain pilots, which often amount to little more than a database upgrade.
Clearing and settlement is a difficult problem that involves the collaboration of many financial institutions as well as industry-wide technology infrastructure upgrades. Finding the right solution might be easier if we stop calling everything a blockchain.
This column does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.
There's at least one big difference. Updating the Bitcoin blockchain involves a computationally intensive process known as proof-of-work. The computations are intentionally wasteful to serve as a signal for the participants' honesty and to allow the system to be open. Accenture says its prototype uses a different so-called consensus mechanism designed for a closed network with trusted participants.
To contact the author of this story:
Elaine Ou at firstname.lastname@example.org
To contact the editor responsible for this story:
Mark Whitehouse at email@example.com