Blockchain powered stock market rebuild started in 2017 delayed again

The Australian Securities Exchange (ASX) is attempting to replace its core trading systems with a blockchain-powered alternative – an effort often touted as one of the world's most significant blockchain implementations. Unfortunately, the project has struck trouble, again. The application in question is called "CHESS" – the …

[ShadowSystems]

Too little too late?

Will it be a TLTL situation that sees the investors deciding to stop flushing money down the toilet?

Oh who the hell am I kidding, they'll keep shoveling the cash down the crapper as long as some sales & marketing wonk keeps blinding them with bullshit, filling their ears with buzzword bingo bollocks, & slipping the occaisional FatBrownEnvelope into someone important's hand.

*Reaches for my Dried Frog Pills in the hopes it can stave off another bout of frustration-induced insanity*

[Joe W]

Re: Too little too late?

Ah, you have been in projects to move on from those mainframe architectures to something perceived more modern - whatever that means. To my understanding (and I am not at the coalface of the migration project) it is pretty difficult to match the high throughput of that "ancient" architecture. Plus the hardware costs we have now will be basically replaced by database license fees ('cause management likes them big #ff0000 players, I guess), and then there's the cost of the project alltogehter... on the other hand, some of the former mainframe big players are reducing their effort, and since one is de facto locked in to a single vendor, there is some heavy price gouging from their part. So in order to have a future you might have to do these migration projects to "modern" (ehr...) systems, what is essentially a spruced up 80x86.

I do not really want to comment on the madness that is blockchain (though on a first glance it sounds compelling - see Friday's article on "bad ideas"). My first pavlovian reflex is of course ridicule.

[hittitezombie]

Still looking for a problem to solve?

Looks like blockchain shills are still trying to find a single problem that they can solve, apart from money laundering and Ponzi schemes...

Good luck to them!

[Korev]

Re: Still looking for a problem to solve?

You forget collecting extorting money using ransomeware...

[Snake]

Re: Still looking for a problem to solve?

The cynic in me is beginning to believe that the major supporters of blockchain "technology" is actually the hardware manufacturers, as they will be making out like a bandit as the world's compute and storage requirements skyrocket into the stratosphere as all the tech quacks try to force blockchain everywhere...in the hopes that they can grab some money from the resulting implementations.

[Ian Johnston]

Re: Still looking for a problem to solve?

Looks like blockchain shills are still trying to find a single problem that they can solve, apart from money laundering and Ponzi schemes...

What do you mean, "apart from". This is a stock exchange we're talking about ...

[wolfetone]

"The Register understands that is a reference to Digital Asset – a startup that champions the Digital Asset Modelling Language (DAML) ASX is using on the project. "

So they're swapping out a language where it's main knowledge base are dying (not the language itself), and swapping it with a language that's not widely used or really standard?

I mean, that can only ever end well.

[ColinPa]

The platform itself – and blockchain – is said to be "performing well."

I recognise these words. They are usually used during basic functionality, running a single task.

It they had said, "When running this at expected peak workload + 50%, with x thousand concurrent users" it performed well. I would be more impressed.

The problem with large systems come down to concurrency and locking/latching

[DS999]

Wrong place for blockchain

The whole point of blockchain is to decentralize record keeping for transactions between parties that have no reason to trust each other. A stock market requires centralized control to enforce regulations, it cannot allow off book trading or trading by unknown anonymous parties.

There's no benefit to using blockchain in a stock market. A stock market depends on trust, so there's no benefit from eliminating that requirement for trading, but the real problem is you'd need extra work to prevent things like someone anonymously buying up all the shares of a company to attempt a hostile takeover.

That's probably why they can't make it work, they have some true believer in charge who wants to turn the stock market into some sort of libertarian wet dream of a stock market without regulation, and he's fighting the people that are trying to make its implementation fit existing law.

Without regulation, there would be no way anyone on the outside could trust any of the information reported about a stock. The price, the volume, short interest, all could be manipulated at the whim of anyone able to add their own self dealing "trades" to the blockchain. Without accurate info, it is worse than the casino it too often resembles today. It would be like a casino where after you bet $1 million on black and the wheel is spun only then are you able to see it has only red numbers!

[Disgusted Of Tunbridge Wells]

Re: Wrong place for blockchain

What's any of that got to do with this?

A blockchain can do whatever you want. Including preventing anybody from adding to it except the regulatory authority.

Bitcoin <> Blockchain.

[Roland6]

Re: Wrong place for blockchain

>There's no benefit to using blockchain in a stock market.

I would suggest there is every reason to suggest that there are many negatives to using blockchain in a stock market...

The whole point of blockchain is to decentralize record keeping for transactions between parties that have no reason to trust each other.

This suggests effectively the stock market replaces central transaction processing with distributed transaction processing, which given the volumes and timing critical nature of the transactions, suggest things can only end badly...

[druck]

April 1st already?

Blockchain in COBOL on Itanium - really?

[Michael Wojcik]

Re: April 1st already?

Blockchain in COBOL on Itanium - really?

No, not really. The new system is not written in COBOL, according to the article. And while it's not entirely clear to me (I may have missed something), I suspect it's not running on HP Itanium hardware, either.

Not that it would be hard to do blockchain in COBOL. Blockchain is just a degenerate Nagle tree where you periodically prune all but one branch using a half-assed consensus mechanism. All you need to implement a Nagle tree is a cryptographic hash. And while it's a bit of a pain implementing a cryptographic hash in standard pre-2002 COBOL (I know, because I implemented MD5 in COBOL-85), that's not what any sane person would do anyway; they'd call a library function to do the actual hashing.

[druck]

Re: April 1st already?

Think not of what may be possible, but the mind-blowing combination of something utterly pointless in a decrepit language on a defunct processor.

[Anonymous Coward]

A delayed software update that adds features to the platform is the culprit.

Who would ever have expected that?

Prototype app outperforms and outlasts outsourced production version

Intriguingly, my lad* was one of blockchain's greatest ridiculers in the days he built his own custom gaming rigs and pirated Windows (he was young). When he became a financial consultant, he suddenly became one of its greatest evangelists. What is is about beancounters?

*AC for his sake.

[Arthur the cat]

Re: A delayed software update that adds features to the platform is the culprit.

When he became a financial consultant, he suddenly became one of its greatest evangelists. What is is about beancounters?

Consultants rather than bean counters I suspect. As the Demotivator poster says, if you're not part of the solution, there's good money to be made prolonging the problem.

[Howard Sway]

CHESS replacement delay

"There's more work needed than we thought there'd be, you're gonna need to write a bigger check, mate!"

This sounds like the project from hell to be working on, hope the programmers are being very well paid indeed for their suffering.

[hittitezombie]

Re: CHESS replacement delay

Interviewer: "We see that there's a gap in your CV for three years, can you explain this?"

Programmer: "I was.. I was.... I WAS IN PRISON! Yes! I definitely wasn't working on a blockchain project, no no no, whoever saw me committing code to that repository is a liar and I'm ready to protect my honour to the death!"

[Persona]

Immutable record

with all entries immutably recorded – just the way traders and regulators like it

The blockchain is not adding tangible value. Financial trading relies on trust. Both counterparties record the trade. When there is a dispute they each examine their records and decide if they or the counterparty is at fault. In the trading cases where blockchain immutability could be of any value it is very very rare for one or the other not to agree it was at fault and any such disputes are readily settled as the firms rely on being a trusted entity. If they lose the trust of other market participants they won't be around much longer.

[G Mac]

Re: Immutable record

Actually, if my memory serves, what happens when something fishy/buggy/something folks don't like occurs in that immutable record is that they fork the blockchain.

"Sorry mate, that stock is on the old blockchain and is now worthless."

[iron]

Re: Immutable record

Or if a majority of nodes agree they just change the "immutable" blockchain.

[Michael Wojcik]

Re: Immutable record

As Bruce Schneier has pointed out multiple times, we have plenty of protocols for append-only ledgers. Blockchain is just a child's version of a Merkle tree,¹ doing a job that can be done in better ways for most use cases.

There are good uses for Merkle trees (and other Merkle graphs). git is popular, for example. I've yet to see a convincing use case for blockchain. This doesn't look like one either.

¹Just realized that in other posts I've written "Nagle trees", having somehow mixed up cryptographer Ralph Merkle with John Nagle of TCP's Nagle Algorithm fame. Oh the shame. Comes from working in too many technical domains, I guess.

[HildyJ]

Si non confectus, non reficiat*

Maybe upgrade your CPUs, but other than that, if it ain't broke, don't fix it.

Nowhere can I find any indication that the current system is not up to the task. This seems like a consultant driven cake with a blockchain icing.

* Lord Vetinari's motto, so you know it must be right.

[Michael Wojcik]

Re: Si non confectus, non reficiat*

Upgrading Itanium CPUs is ... tricky, these days.

But, yeah, we'd be happy to sell them a COBOL runtime for shiny new x86-64 blades running Linux. Or whatever their beverage of choice might be.