Modernising retail banks’ infrastructure is so time consuming and expensive that many will look to migrate customers to new digital subsidiaries, according to Tim Hooley, chief technologist, EMEA Financial Services at Red Hat.
“If you have monolithic software you cannot be innovative,” said Hooley, who was speaking on the side lines of Sibos in London last week. “They are leaving their old banks in place, with their old software, and most of them are hoping I think that their customers - and you’ll have to forgive me for saying this - will either die off, or open new accounts with their new banks.
“They are never going to upgrade their old systems because they are too expensive. Their data centres are hugely expensive, they’ve got thousands of developers, they’ve got thousands of operations people doing the data reconciliation between all of the different data sources that they’ve got.”
Many banks previously built applications in a monolithic way. This would require one giant data set of code to be written and created a high dependency on those that wrote the code to be able to make appropriate changes in the future. But the introduction of microservices-based applications has meant that applications can be broken up into the smaller units. A core banking system may break up into 200 pieces, meaning functions are more easily changed independent of the rest of the system.
Hooley has had several conversations with leaders of fintech organisations who are similarly bullish on microservices.
“The Thought Machine CEO was telling us that he thinks that banks will get to the point where they pay their customers to move to their new banks…so that they don’t have to migrate their old systems to new systems,” said Hooley. “[He] said that the cost of having a customer on a current account on a new bank with a microservices based platform versus having a customer on an old bank - the difference in cost is 100 times.”
A lack of skills around microservices requirements is also holding banks back from making the change.
“If you are a developer and you have been at a bank for twenty years, you don’t know how to do microservices – it is a completely different way of building software,” said Hooley. “There are millions of lines of code that nobody knows how to change because the people who wrote it have left, and they didn’t document it very well. If you change one little bit of it you have to test the whole thing, which makes it really expensive to test.”
But the monolithic method of code writing is also cause for concern when selecting a software vendor, according to Hooley.
“That is an interesting problem because if you use a software vendor - which most people do - their software is largely built the old way. If you are a customer of them then you are in trouble because you can only change your service to your customer as fast as they change their software, and they can’t change their software very quickly because it’s built the old fashioned way,” he said.
But for Paul Camp, chief executive officer of treasury services at BNY Mellon there is a need to invest in current systems that are used daily by clients while also look to the future.
“What we are very cognisant of - and the technology companies don’t always get it - that our clients need both,” said Camp on the side lines of the conference. “They need the stuff which works today and has worked for years, because that model is not going to switch instantaneously, and they need a path to the future.”