I came to NCR to sell a modern ATM driving and transaction processing platform. The goal was to leverage a platform that was focused on transforming messages and building in flexibility to adjust for what was to come, not what was already on the shelf. Part of that was thinking about how we formatted transactions but also how we got the most out of the devices where these transactions started. NCR the premier provider of ATMs has to consider the past as do we all and that sometimes gets in the way. So modernize the middle to get to the future is a pretty decent approach.
ATMs mostly run on legacy platforms that use ISO 8583 or derivatives in the NDC spec; and this is due to legacy connections directly to device components. As we start to reconsider how this works – the interesting approach we were taking was to translate these messages in the middleware and that worked. But the company considered improving the device driving as well building a new Terminal Handler to update the protocol itself the device uses. That’s great but change is hard and with so many decisions it became unmanageable.
The core integration also became unmanageable – we had so many ways to do this, as well as leveraging legacy network connections that we continue to churn in market. The reality is all of the solutions we offer, from web services custom interfaces, to microservices based flexible mapping tools to legacy mapping tools built into the switch – all work. But change is hard and so is strategy and how to get there is often driven by legacy thinking and confusion on the future end game. Do we expect cores to persist? Do we instead build middleware? Do we need transaction processing and to own technology? Do we outsource to ATMaaS (as a service) providers and just give up?
The ISO 20022 spec has become more mature as of late. With consideration for Bitmap legacy formats to ISO/XML based messages – the latter is really the way to go. And this is true for other systems as well. Wire and Faster Payments are headed that way. TCH RTP is ISO 20022 based, and FedNow will be. I’ve been working with XML based messages for years in web banking technology and its has all but replaced legacy BAI file based messaging; but they persist as well.
Finally its all moving well to the cloud – deploying in containers, as a big old blob or as true microservices – the cloud service providers (CSP) have removed roadblocks of big iron. Smaller FIs can invest in resources that use these technologies and run with CSP based operating environments – faster, continuously innovating, in partnership with modern vendors.
I’ve moved on to Digital Banking and Microservices based common business services that enable cross channel, cross device workflows. And I’ve built strategies with customers looking to execute on this. But the payments and transactions persist and should be a part of these discussions. I see it coming and look forward to it. Bringing these new capabilities will mean streaming data, immediate not just real-time transaction processing, continuously connected and interactive services with shared sessions between clients and providers, even clients and AI. Here we go