I heard Ripple is trying to market for this use case - Foreign Exchange transactions. Unless I've misunderstood it sounds like HSBC are more interested in the forensics of the transaction rather than having a unitary value for each transaction (which the value of Ripple would dictate).
HSBC suggests it might have found a... use for blockchain?
HSBC claims to have settled three million foreign exchange (FX) transactions and made payments worth $250bn using distributed ledger technology (DLT). The bank said it had made "significant efficiencies" while using its DLT product, HSBC FX Everywhere, for the past year – suggesting the risk-averse financial sector is treating …
COMMENTS
-
Tuesday 15th January 2019 13:16 GMT A Non e-mouse
Truth
shared, single version of the truth of intra-company trades
Because blockchains are immune to people fiddling the system, aren't they...
-
Tuesday 15th January 2019 15:50 GMT Lee D
Re: Truth
"To execute the attack, the miner acquired at least 51 percent of the network’s total hashpower, which provided them with temporary control of the blockchain"
Sure, if one malicious entity is able to enter arbitrary transactions against your network without your knowledge, after taking control of half the people able to enter transactions on said network (note, this wouldn't include *users* or people *viewing* the chain to verify transactions, only those capable of making their own).
I think you'll find that any distributed ledger would be vulnerable to attack in such a circumstance.
You're literally asking "half-the-world's foreign-exchange banks" to collude to make a transaction that the others will accept as genuine, in a way that's blindingly obvious, flags all kinds of warnings immediately, and would only proceed if you were stupid enough to not put decent checks in your system and wait for the right number of verifications from trusted entities - at that point, you have bigger problems than what you use to record that transaction.
-
Tuesday 15th January 2019 16:32 GMT FrogsAndChips
Re: Truth
That sounds exactly like how Kerviel managed to engage 50 billions in DAX futures - knowing how to bypass weak controls and convincing people to ignore red flags. Granted, that was 12 years ago and banks have learnt from this, but we all know how history is doomed to repeat itself...
-
Tuesday 15th January 2019 16:34 GMT Persona
Re: Truth
"I think you'll find that any distributed ledger would be vulnerable to attack in such a circumstance."
As it was intra-company I think that you will find HSBC's ledger (which might have been distributed but probably wasn't) was handled solely within HSBC's systems which voids the concept of a 51% attack.
-
Wednesday 16th January 2019 15:59 GMT Michael Wojcik
Re: Truth
I think you'll find that any distributed ledger would be vulnerable to attack in such a circumstance.
This is precisely the problem, and why BGP attacks attempting to partition the Bitcoin mining network happen around a hundred times a month.
"Such a circumstance" turns out to be quite common.
-
-
-
Tuesday 15th January 2019 13:26 GMT katrinab
Just to be clear
Buying Bitcoins won't give you exposure to this, because anyone can write their own blockchain, and HSBC presumably have.
Ultimately, it is just a way of storing data, just like an Oracle database or an Excel spreadsheet are ways of recording data. These things are of interest to techies like the people who read this website, but not really that important in the overall scheme of things.
-
Tuesday 15th January 2019 16:49 GMT gosseyn0
Re: Just to be clear
The small and private blockchains such as the one HSBC is using can easily be hacked as we've seen on numerous occasions. It's not inconceivable that they instead use the Bitcoin blockchain since it's the biggest, oldest and most secure blockchain in existence. It could be implemented as a side chain with all the data being encrypted. Thoughts?
-
Wednesday 16th January 2019 10:53 GMT Colin Bull 1
Re: Just to be clear
"These things are of interest to techies like the people who read this website, but not really that important in the overall scheme of things."
I think the following paragraph might create a step change in the financial system.
"It transforms the process around intra-company foreign exchange activity, automating several manual procedures and reducing reliance on external settlement networks."
At present credit and debit cards are an effective duopoly of Visa and Mastercard, both charging exhorbitant fees. This could eliminate these parties skimming every transaction
-
Thursday 17th January 2019 17:43 GMT katrinab
Re: Just to be clear
"automating several manual procedures" is something computers do. You don't need a blockchain to do it. Any sort of computer program will do.
"reducing reliance on external settlement networks", by replacing it by one run by HSBC. Do you think HSBC will be any cheaper than Visa or Mastercard? And again, HSBC could have an internal network that stores its data in a different format[1].
[1] Response in advance to anyone who replies telling me why Blockchains are so amazing:
You are probably correct, but, most of the problems that Blockchains seek to solve are not problems that exist in reality. Most of the fraud examples you see in real life, Bitcoin wouldn't stop. For banks, the risks are that someone steals a customers credentials, they could just as easily obtain the relevant key for their wallet; or loan application fraud - the decision to lend or not lend money has nothing at all to do with blockchains.
Blochchains would stop someone accessing the bank's ledgers and changing the recorded transactions, but this is not something that ever happens.
-
-
-
-
Tuesday 15th January 2019 14:36 GMT John Sager
Re: immutability provided by DLT
Well, we don't know how HSBC have implemented their tech. I would be very surprised if they used a 'proof of work' system like Bitcoin & Ethereum. Would they really want to spend that much on electricity? As it's a single organisation they don't necessarily need the 'double spend' protection that Bitcoin needs, so there will be other ways of giving the process the immutability guarantees required.
-
Tuesday 15th January 2019 15:54 GMT Lee D
Re: immutability provided by DLT
At a minimum, I'd expect that they'd need 51% of co-operating foreign exchanges to recognise the transaction as valid before it could be "falsified" into the blockchain.
Coin-based blockchain attacks are when malicious users own 51% of your blockchain AND you choose to recognise their branch of the blockchain as definitive.
There's two different things at play, and a distributed ledger is going to have a lot more control stopping you getting 51% of the ledger under your control - maybe even by the simple precept of "there are only a hundred master nodes and HSBC control 20, another bank controls another 20, etc.)
-
-
Tuesday 15th January 2019 14:22 GMT James Anderson
Risk Averse?
My experience of merchant banking is they are quite the opposite of "Risk Averse".
Certainly in their financial dealings they will take any amount of risk if it generates larger bonuses.
Their IT departments has a penchant for shiny new technologies and bleeding edge software.
The first time I saw a serious Linux box deployed for a serious application was at an investment bank in 2003.
-
Tuesday 15th January 2019 15:40 GMT FrogsAndChips
Re: Risk Averse?
IB are more and more risk averse. They now prefer the safe and steady revenue streams of sales and arbitrage to the risks of trading for their own. Plus the regulatory constraints make it more and more costly to perform their activities. A lot of IB have been shutting down or reducing their prop trading activities in the last years.
-
-
Tuesday 15th January 2019 16:06 GMT Lee D
Re: Probably different model
Absolutely.
Bitcoins are basically generated by an anti-spam measure to stop one person hoarding all the coins without a proportional expense/effort on their part. You could have literally let people get a Bitcoin just by sending an email to an automated address, if you'd wanted, and used that as the ledger entry.
A ledger itself, though, just needs to know who made the transaction, who signed off on it, who can revoke it, and then put it into the ledger. Importantly, even complete compromise of their key cannot erase the history of what that key did, as each transaction is in effect "signed" by the next (basically the core idea of blockchain).
Think of it not like Bitcoin but like those play-by-mail games of old. You collect the transactions, perform the calculations, and modify the "database" in the middle. Then you send out the results of the next "round", everyone else can see the effects of those new results and "previous moves" but you can't change them.
The only difference is that instead of one person performing the calculations, you're publishing the data, letting everyone do the calculations at the same time, waiting for everyone to agree what the result is, and then having everyone sign off on that result.
Bitcoin does that, WHILE solving a pointless maths problems that's hard to solve (I believe it is literally "hashing the entire existing blockchain with random numbers until the hash results in 00000000000000000000000" - or it certainly was at one point). It's literally a time-effort-and-money-burning exercise so that there's something that has to happen to make a coin.
-
-
Tuesday 15th January 2019 17:01 GMT colinb
$250,000 are you sure?
This is barely a rounding error on the rounding error in banks.
Other sources like the FT has it as
3m FX transactions worth $250bn thats billioooooon
I tend to read these puffs pieces as: our internal tech is shit, really shit, we done this other thing that that everyone is talking about and is less shit.
Aren't we great.