Historically, payment processing in financial organizations has been implemented as very slow, batch-file oriented processes, resulting in electronic transfers taking often, multiple days.
However in recent times there has been a push for modernization of payment protocols towards the implementation of real time payment by many financial organizations.
Most recently this includes the Federal Reserve, which has released FedNow which is their solution of a real time payment service.
The Clearing House has also released it's own real time payment system: RTP
And of course in the US, the well known Zelle platform which is primarily consumer-focused.
For organizations that wish to leverage real time payments, this is often a significant technological challenge, specially for those that already have a significant legacy built around daily batch processing.
Mulesoft is a perfect choice to implement such an integration because while it's well known to be a powerful platform to build REST APIs, it's also just as good for the implementation of real time or near real time systems.
This includes support for integrating with many different messaging systems (like ActiveMQ, Kafka, AnypointMQ, AWS SQS and many others).
Documentation isn't neglected either, as support for the AsyncAPI standard is incorporated in the Anypoint Exchange platform, allowing to clearly document all data flowing through the system with a high level of quality.
In addition to all the the connectors provided by Mulesoft, it also provides the Mulesoft accelerator for financial services, which provides templates for various use cases which greatly helps teams getting started with finance integrations, although it doesn't include real time payment integrations at the time of writing this article, it does support ACH for example.
Integral to a real time payment solution is the buildout of a composable architecture. Pluggability and loose coupling to ticketing systems, banking core systems, and gateways will allow for greater ease in the case of a change in end-points.
The following diagram provides an insight into the high-level logical architecture for a payment hub. While not a one-size-fits-all solution, this diagram shows the major components of the architecture, while still offering the ability to modify these components to fit your business needs.
First, it has a very strong focus on incident detection and management. It can detect when something goes wrong and a transaction doesn't completes as expected within a certain timeframe.
It provides integration with ticketing systems as well as a diagnostics UI allowing the appropriate staff the ability to be notified when unexpected incidents occur requiring human intervention. Payment Hub also includes the ability to gather all relevant information ( message payloads, logs from all systems relevant to that transaction ), as well as the ability to modify and re-submit messages for processing.
This architecture is designed to support stream processing. For example, you can easily incorporate capabilities like fraud detection, analytics, etc.
Syntek also provides our own accelerator asset to help our financial services clients quickly implement payment hubs that support multiple types of payment systems, including real time ones.
Note that there are some fundamental differences between the MuleSoft Financial Services accelerator and the Syntek Payment Hub providing some significant benefits.