Bitcoin Forum

Why do bank IT systems keep failing?

http://www.theguardian.com/money/2014/jan/27/bank-it-systems-keep-failing-lloyds-rbs-natwest (http://www.theguardian.com/money/2014/jan/27/bank-it-systems-keep-failing-lloyds-rbs-natwest)

"Lloyds, RBS and NatWest customers have all been locked out of ATMs or unable to use their cards at check-outs. A lack of investment is only compounding the problem."

Ah - music to our ears :)

Generally bank systems are very poorly built in a rush with exceptions to architectural advice all over, this is piled on top of older systems equally bad going back to some systems so old they beg staff not to retire and scour ebay for parts.

Bank maintenance is outsourced to the lowest bidder or the ones with the most hair gel.  The outsourcers can then only make a profit if there are faults.  The result is pretty obvious: the systems develop faults. 

It's a lot like jenga: push a bit here, pull a bit there and over it goes and no one quite knows why ;). (But the boss buys everyone beers and pizza when it happens 'cause it's just paid everyone's wages for the next few months  :D )

I program for a bank and this is 100% spot on. Another point is that legislation changes in the finance sector are non-negotiable and often code must be upgraded by specific dates set by the government. So this throws a spanner in the works for bug fix and clean-up projects...

Because IT systems are programmed by people, and people make mistakes.

+1

Over the years I have seen so many senior managers blaming 'IT failures' for a major mistake rather than admit their own incompetence that I remain skeptical whenever I hear about these problems.  IT problems are a scapegoat that often passes by without any rigorous examination.  

If a major high street bank had a liquidity problem that was preventing them from authorizing payments, wouldn't it be easier to just blame IT systems and try to resolve it the next day rather than admit the truth?

If I were to put on my tinfoil hat I'd say that this is something the banks are doing to make it more and more difficult to get your money back and their back end systems will work perfectly so when another financial crisis comes along they can sneak off with all the money :P

People often compare programming to bridge-building and complain that the programmers suck.  But really, it should be compared to building bridges where every day thousands of brigands come up the river with all manner of weapons trying to destroy the bridge.  And it only takes one to succeed to make big headlines.  And sadly, the programmers actually could do it if their bosses would listen to them and give them the time to make the bridge correctly, but it would take an extra 2-3 weeks, so no.

Halon said "Never attribute to malice that which is adequately explained by stupidity." But he's wrong when it comes to banking systems. I think when you say "mistakes" it should similarly be in quotation marks.  Nearly every one of these faults I've looked at were deliberately built in.  Sometimes you even read comments where the programmer is off with a stream of profanities about how they were told to do this that or t'other against their better judgement.  And they were right.  ;D

Yes, my brother-in-law works programming satellites and there is no end of "mistakes" that creep in, mostly because of management interference in one way or another.

Because like most big businesses they have a lack of investment and tend to be built on over the years.

Ive worked in systems development for 30 years and have seen it 1st hand.

A lot of boards, particularly FDs see IT as a necessary evil and dont invest in new systems, they simply bolt on new features with the result that the systems mushroom and become frankenstein monsters which are difficult to maintain and develop further.

Just my 2 pennies.

Very much this.  I know of a company running a DOS/Java/Oracle clusterfuck as their main platform to serve a very large customer base.  And up until recently I'm pretty sure my Bank were running a DOS program too.  It's a bit scary.

Because banks are complex beast and their IT reflect that.  Most are running 80's mainframes for core services because they are reliable.  Laugh at that, but if you know the number of cockups on non-mainframe systems for the same transaction volume, you'd agree. Or even on mainframes, my company runs major data collection and processing across the world, similar to bank scale, but the data is fudged and sticky plastered or pushed out late, no one notices or an analyst kicks up a fuss and gets their next report free.  Bank does that once and its a major event.  Of course, outsourcing operations to third parties and losing hundreds of man-years of experience is their own doing (the case with RBS)

We'll all learn this in a few years time as Bitcoin nodes are running multi-terabyte blockchains.

To upgrade the systems on any major bank would be a major up-taking.  I've worked for 2 large banks in the US and the computer systems were at least 10 years outdated both times.  And when they do update, its patchwork upgrades.  Banks are under so much pressure these days to increase earnings to please the stockholders, no room for internal computer updates.

... these massively centralised systems are doomed to failure, it is a mathematical surety.

We just have to prepare for the ultimate collapse when the failures cascade from one megabank to another, and finally the interlinked SWIFT, SEPA and FedWire settlement systems crash simultaneously.

Bank IT systems are usually legacy systems, built over existing legacy systems with hundreds of patches applied.  When it takes years to develop a new piece of software you're hardly going to throw it away when Shiny New (tm) comes along in the IT world.  My local branch still runs Windows XP on all branch computers.  It also means the layers and layers of patches and hacks break down.  Often.  That's why IT systems keep failing.

Bitcoin isn't the answer either.  You're either looking to store your bitcoins on someone else's computer (hacks, theft, government confiscation risk, etc) or if bitcoin becomes useful in daily life we'd quickly be looking at a petabyte blockchain.  Whoever creates bitcoin v2 without the exponential growth of the blockchain will create real value.  Dogecoin can die in a fire.

Kind of reminds me of analogue TV signal vs. digital TV signal...

Some parts of the world still operate on analogue only. TV picture is the same as it ever was.

Some parts of the world are in switchover status, with digital and analogue being broadcast simultaneously, and the analogue scheduled for mothballs. The analogue TV picture in these places is intermittent, sometimes it's downright unwatchable. Instead of saying "we're not putting much effort or resources into the maintenance of analogue, it's turning off soon anyway", the excuse is "these old systems, they get more complicated to maintain as time goes on, can never find all these outdated parts cheap enough" etc. Both sound believable, yet the most convenient excuse is used. Meanwhile, analogue TV is still working fine where it's the only option.

It's this costly infrastructure that the banks hate to invest in that bitcoin and it's community have to compete with.

I don't have enough technical knowledge to understand if bitcoin will scale to process the amount of transactions banks do, but if it doesn't, I'm confident an alt coin will be able to figure out a solution.