Last week at the DevCon1 conference, Lee Braine of the Investment Bank Chief Technology Office at Barclays challenged a hall full of Ethereum developers to consider how smart contracts could cope with some of the complexities of investment banking processes – and mentioned securities settlement as an illustrative example.

It was an interesting mix of cultures at the conference. The be-suited Dr Braine has 20-plus years of banking experience. He was joined on a panel to discuss banking and the blockchain by Ian Cusden of the R&D team in the Chief Technology Office at UBS, who has been using Ethereum to prototype issuing smart bonds in a sandbox environment at Canary Wharf's Level39. Also on the panel was Rex Maudsley of Société Générale's investment banking architecture function, who has been exploring innovations and collaborations with FinTechs on technologies such as blockchain. Andrew Keys and James Slazas of ConsenSys were also on the panel.

Before them was an audience of leading decentralised app developers and hardcore Ethereum cypherpunks; some wearing dreadlocks sat cross-legged on the floor with laptops in front of them.

Braine began by saying he has been assessing blockchain-related technologies in the past year from an architecture perspective, specifically looking at interbank shared ledgers and executable business agreements, including technologies such as smart contracts. He said that many investment banks are aiming to reduce costs where appropriate and that both shared ledgers and smart contracts may have the potential to help. But he emphasised that there are existing stable and reliable infrastructures in place, including managed networks and utilities, and that enhancing or integrating with or providing similar functionality can be complex.

Braine's challenge went like this: "I'm really throwing it out to the community to propose innovative technology solutions. Many of the smart contracts you currently see posed by people are designed to solve a problem on something like Ethereum, but bear in mind when you consider models where you've got, for example, securities settlement, FX settlement, etc, that there are netting processes. There are processes and workflows that are a lot more complex than small smart contracts. For example, transactions can be pooled together into a netting pool – and the netting processes can be much more complicated than simply adding up credits and debits via summation.

"Netting optimisations can include trying lots of different settlement combinations in order to achieve a high settlement efficiency. So how do you do that? What you can do is effectively 'stop time' for a short period: take a snapshot of transactions from the banks, perform the complex optimisations which may be equivalent to scenario modelling and then, at the end, release the optimised results back to the banks.

"Incorporating these sorts of complex challenges into smart contracts in some way will be hard. I think such challenges would benefit from the focus of not just banking and FinTech, but also input from academia as well. So, one of the things we are looking at is collaboratively shaping some of the really tough technology challenges for smart contracts and engaging across the board with banks, vendors, FinTech, incumbents, academics, etc on experimenting and solutioning. I think we'll know that smart contracts technology has fully come of age when it's able to efficiently model processes as complex as netting optimisation."

Andrew Keys of ConsenSys, who also has a banking background having worked at UBS, acknowledged the consortium approaches and the request for standardisation, but pointed out that banks also now have access to tools to experiment with the technology: "Rather than define standards, give them the ability to understand what this technology is, which is somewhat pursuant to what we are doing now with Microsoft where we give this test bed for the banks to play. So, with that said, what do you think the future will hold?"

Braine answered: "You mentioned one particular avenue of experimentation. If we look at multiple avenues, then we can actually see them cross-pollinating.

"Effectively you have experimentation: you have Ethereum and other technologies that we can play with, we can experiment, and we can explore the functional and non-functional behaviours.

"But in parallel with that, we also need a structured design method: architecture, design, and engineering. This is a common theme that we have been encountering over the past six months or so, that it's necessary to interplay both of these avenues so that the two can learn from each other.

"I think one of the things we found in the last few months is the incredible speed at which certain technologies have matured. When we were looking at some of the experimental technologies about nine months ago, many would fail at the initial hurdles if we were assessing by data privacy, scalability, performance, etc. It's encouraging to see, as the months have gone by, how start-ups are pivoting, making adjustments to their propositions and technologies in order to come back with things that are more suitable for investment banking's needs."