2010: what’s in store for the technology of banking?

After various predictions yesterday, including the return of Glass-Steagall and a Chinese bank buying an American or European one, today it’s the turn of technology.

There are lots of questions about technology, such as:

• Will Google make us stupid?
• Will we live in the cloud or the desktop?
• Will social relations get better?
• Will the state of reading and writing be improved?
• Will online anonymity continue?
• Will the Semantic Web have an impact?

Lots of things we could talk about, but here are five key things that will happen this year with banking tech:

• Banks will start using cloud computing services across the board
• Real-time becomes normal
• Real-time disrupts the status quo
• Mobile services will continue to rip through banking and payments innovations
• Contactless payments will finally start to make some inroads

Banks will start using cloud computing services across the board

The use of cloud was a big discussion point through 2009, with this blog arguing that many bank infrastructures could move to cloud style services if they wanted to. But then some say cloud is just old-style mainframe computing, except that it has moved from the host contorleled environment beign private in-house to being public and open. Whatever.

The fact is that today’s investment markets, due to latency requirements, are moving towards cloud for trade execution regardless. Proximity hosted, co-located servers will sit in massive server farms in Chicago, New York, London, Frankfurt, Mumbai, Shanghai, Hong Kong, Tokyo, Sao Paolo and other major trading cities, all connected by lightspeed communications enabling global trading to take place faster than the blink of an eye. These comms lines already allow a global packet of trade data to move 32,000 kilometres around the world in less than 400 milliseconds (that’s two really fast blinks of your eye) and networking firms tell me they will get these speeds down to under a millisecond by the end of this year/early next.

Therefore, the idea of having everything available through the Web on a secure plug and play connections anywhere, anytime (including on your mobile) will be the way to architect throughout 2010.

Real-time becomes normal

With such lightning speeds, the idea of connecting and then having to wait three days or more for something to happen will become so last century. Banks that sit on funds for days will become targeted by consumers as weak banks. I mean, when you use a mobile phone to make a call, is it acceptable for the call to be answered in D+3? Send a tweet and get it 24 hours later? No way.

How long do you take to answer an email or get annoyed that someone hasn’t responded to one (oh yes, and email is also way out of date – use Facebook and Twitter instead)?

So the idea of D+ or T+ is going to die a death in the 2010s, starting right now. Payments, transactions, cross-border money movements, fraud checks ... everything in banking will move to real-time.

The real-time infrastructures will change the nature of banking and are of concern to some – “if we saw Lehmans collapse over weeks, what would happen if they could collapse in real-time” – but the movement to real-time is inevitable and unstoppable.

However, the thing that goes with real-time money is real-time risk and real-time risk management. This is why real-time analytics and real-time reporting will become key. For example, the UK’s FSA has already required trading firms in London to move towards real-time trade reporting, a move that will supposedly cost upwards of £2 billion to implement.

Real-time disrupts the status quo

Now, real-time doesn’t necessarily mean that you give the customer real-time. You might charge for real-time or near real-time, and that’s where it gets interesting. After all, if you can process a payment of £100 million for the same price and in the same speed as a £1 payment, then where is the bank adding value?

Equally, infrastructures that inhibit real-time – such as the complexity of SWIFT messaging around transactions – may start to call into question the raison d’être of those infrastructures.

SWIFT has been challenged many times over its raison d’être – Heidi Miller, 2004, “if the internet is ubiquitous and free, why should we pay SWIFT for messaging?” – and they’ve continued to innovate and push for change through SWIFTNet and XML messaging.

Now however, the community based infrastructures of the past – created due to the cost barriers of implementing technology in the 1970s on a global scale – are no longer necessary. This si why Visa and MasterCard IPO’d, and why community infrastructures will give way to open infrastructures with secure networking layers.

This last point will be a critical challenge to SWIFT and others – the London Stock Exchange, for example – in 2010, and will see further upset of the status quo.

Mobile services will continue to rip through banking and payments innovations

One of my favourite subjects of 2009 was the mobile telephone as an interface to the bank. This was particularly down to the fact that 2009 was the year that mobile internet become easy thanks to the iPhone.

The iPhone will transform banking generally, and is already proving to be a killer app for finance thanks to iPhone apps.

iPhone apps abound for finances, http://www.apple.com/iphone/apps-for-everything/managing-money.html with Bloomberg, Mint and more featured. Banks have started rushing out iPhone apps too, with NatWest being one of the first UK banks to major on this http://www.natwest.com/personal/more-ways/g1/mobile-phone-banking/iphone-users.ashx, although BBVA, Bank of America and others have been on the case for a while now.

But the transformation starts to happen when you see innovators like Twitter co-founder Jack Dorsey starting to think about using the iPhone as a payments system.

This is his new venture Square, a simple clip-on to the iPhone headset jack that turns the phone into a payments processing machine:

Now that clip shows a very heavy US-centric view of the payments world and what Jack doesn’t know is that the mag stripe and signature is also so last century.

We’re now into contactless cards and Chip & PIN.

Contactless payments will finally start to make some inroads

For a while, contactless has been discussed and hasn’t really taken off. This is no reflection on the banks, as several have tried to rollout major programmes such as the Chase Blink Card  and Barclaycard OnePulse. Barclaycard recently reached five million cards issued for example, but that still doesn’t make this a major success ... yet.

The reason is not enough places to use a contactless card.

But that’s changing, as can be seen in this video:

Suddenly, bringing together the mobile phone with apps that turn it into a point-of-sale doesn’t seem stupid or strange ... it just makes sense.

Like Jack Dorsey’s Square, turning the mobile into a payments system for both transmitting and receiving money is not the future ... it is here and now.

And that’s what we will see more of in 2010. More innovations on the mobile internet to make money on a mobile as easy as cash in your pocket.

Now that makes 2010 a very exciting year doesn’t it?

Comments

[Tyler Hannan]

I find the perspective given in the "real-time becomes normal" section exceptionally intriguing...

As an avid real-time communication user (and yes, that includes my mobile phone) the assumption of a real-time experience, akin to what I experience in my social endeavours, in my financial experience is a logical leap.

Given my experience with transaction processing and payments in general, I've come to "accept" (albeit grudgingly) the "why" behind the D+ and T+ discussions. But, as a typical consumer, (or even a technology savvy developer with little experience in transactional engagements) the expectation of real-time, put in the context given in this post, is understandable.

It is the process of putting ourselves into the mindset of others where we can gain additional clarity...

[Thomas Barker]

I doubt you see any move to clouds (in-house or public) for trade execution. I know people who work on low latency stuff - banks are starting to use specialised h/ware. Coding stuff into network cards, GPUs for realtime modelling, etc.

No one's going to save $5000 on h/ware for a $1m/sec system.

Mobile phone payments are going to take over from cards fast though... There's just so much more they can do over a card, and you don't need to mail one to each customer.

[Alain Rogister]

This raises two points:
1) No actual lessons have been drawn from the latest financial crisis. We are happily going towards a real-time cluster**** since no one can possibly predict with 100% accuracy the consequences of a fully (near-)real-time global financial system. Wishful thinking on real-time risk management cannot solve this.
2) Since light travels in optical fiber at 2/3 its speed in a vacuum, the lower bound on a 32000 km hop is 160 msec. Could you provide a reference to these wild claims of 1 msec ?

[Chris Skinner]

@Tyler

Thanks. Real-time as real-life is here!

@Thomas

I guess I see colo and prox hosting as moving towards private clouds for trade ex, but maybe that's just putting together 1+1?

@Alain

Agree on point (1).

On point (2), it does amaze me how low latency network providers are pushing the envelope here. THEY tell me that it is possible right now to do a global data hop in under 400 millisecs and, by end of this year, 1 millisec should be possible.

Go figure.

[Alain Rogister]

Chris,

Thanks for your answer. Ultra-low-latency requires proximity: 1 msec places a theoretical upper bound of 200km distance (i.e. tens of km in practice). Don't let vendor talk trick you into assuming that the law of physics are going to change :-)

[Steve]

I firmly believe that real-time payments and settlement processing, in today's global economy and financial markets, will eventually do away with banks having to settle their accounts at end-of-day with their respective central banks, thus allowing banks to process payments freely 24/7 around the globe. Trade transactions will continue to be settled by a certain date/time, but not a bank's entire central bank account. If this happens, the implications are huge, especially in the overnight funding markets. Also no more netting of positions at end-of-day; everything is real-time.

[Chris Skinner]

@Alain

I understand the proximity game and low latency challenges. The original 'limitation' I understood was 1 millisec for every 60 kilometres. I am now told this is being overcome through direct colo links between major centres.

Agree with you that there has to be some data drop even with such proximity hosting and colocation, but interesting to conjecture the possiblity that a real-time global loop could be feasible one day (and maybe one day soon).

@Steve

Wowser - when we get into real-time, a lot of interbank and intrabank processes become questionable or even redundant. Now that'll be the day.

Was fascinated for example by one payments head saying to me: "whatever our IT department say, it does not make sense to process a £50m for the same price as a £5, even thought it costs us no more in processing internally".

Hmmmmmm ...

[Marcel den Hartog]

As we have seen in recent publications, a large Korean credit card processor has moved its business critical applications from distributed platforms to a (new) Mainframe (http://www.theregister.co.uk/2010/01/05/ibm_bc_card_mainframe/). Back to the future? Many large companies are slowly starting to realise that the current infrastructure of distributed systems is expensive, inflexible, not secure and not very green ( http://www.ca.com/Files/SupportingPieces/ca_mainframe_survey_report_208226.pdf) Administration and Management costs are skyrocketing and do not scale against the actual workload of these systems. Many large banks who already own an IBM zSeries Mainframe have already changed the status of their Mainframe from “Obsolete, let’s get rid of it” to, “Strategic, cost efficient, it’s a major part of our future IT Strategy” (http://www.ca.com/Files/SupportingPieces/vertical_findings_executive_summary_215877.pdf) . The reasons are obvious, a Cloud running on multiple virtual zLinux machines that closely connect to > 60% of the corporate data and are fully under control and very secure is not only very attractive, but the costs associated with it will be substantially lower than building an internal Cloud infrastructure. And now we are seeing that companies are even buying new mainframes. With the current state of the art technology and attractive pricing from IBM and ISV’s, more will follow…