With Digital and Instant Payments growing at a massive scale across the globe, FIs (Financial Institutions) with legacy systems keep hearing the biting words as Maverick in Top Gun – “The end is inevitable. Your kind is heading for extinction!!”
Technologists there wish they could rebut as Maverick – “It’s not the plane. It’s the pilot!!”. Unfortunately, Instant Payments or Real-Time Payments processing is a different ball altogether and they must bite the bullet to revamp their payment systems for the future.
The Changing Payments Landscape
The global digital payments market is anticipated to generate a revenue in the range of USD 310 billion to USD 390 by the end of 2030, highlighting its immense potential for financial institutions worldwide. Brazil’s PIX and India’s UPI are two of the most successful modern Instant Payment systems in the world supporting a variety of use cases.
Launched in 2020, PIX now has nearly 800 FIs onboarded as participants and processed 110 million transactions per day on an average in June’ 23. The immense popularity of PIX as a mode of payment in Brazil is such that the total PIX transactions surpassed the combined total of debit and credit card transactions within a year of its launch.
India’s UPI which was launched in 2016, currently has 473 participants and processed around 311 million transactions per day on an average in June’23. According to PWC’s Indian Payments handbook published in May’23; UPI is storming towards clocking 1 billion transactions per day by FY 2026-27.
In markets like the US, cards as a mode of payment have always been a lucrative offering for Banks, which is probably why initiatives to implement similar real-time payment rails have been on the back burner for long in these markets. However, With the debut of the FedNow payments rail in July'23, the US Federal Reserve has certainly taken a cue from the success of the likes of UPI and PIX which continues to inspire and encourage with its successful day-to-day payments that proves to be easier, cheaper, and faster for individuals and businesses alike.
Are your systems ready to endure this payment revolution?
The numbers above are certainly phenomenal and the future promising, however the big question is; are your systems designed to sustain this transformation? That’s the question every Technologist at FIs embracing this change should ask themselves to enable their organizations to capitalize on the opportunities in this space in the forthcoming years.
So, what are the key design attributes that make your Ideal Next-Gen Real-Time Payments processing platform?
Back to the Drawing Board!
Systems processing real-time payments generate, collect and process humungous amounts of data along with complex integrations with multiple other systems. The most common non-functional requirements from FIs and regulators are:
While many architectural aspects contribute to making a perfect system, here are 5 system design attributes that will help you tick some of the above-listed boxes when evaluating or implementing a real-time payment processing system:
1. Horses for Courses - Follow architectural styles that best fit the need.
Real-Time Payment processing demands speed, fault tolerance, and the ability to sustain sudden surges. These are primarily the reasons why most instant payment systems leverage asynchronous messaging. Event-Driven Architecture and Microservices Architecture complement each other seamlessly, resulting in the perfect recipe for building systems responsible for processing large volumes of real-time data. For those unversed with these paradigms, microservices are loosely coupled interfaces that produce or consume events to perform a very specific business operation, and event- driven Architecture refers to a system of many such microservices that exchange information between each other leveraging a messaging backbone.
Microservices design patterns like the Saga pattern for service collaboration, the API Gateway pattern for cross-cutting security concerns, and data transformation requirements cater to many of the standard requirements for payment microservices. With all its benefits, asynchronous processing is complex due to challenges associated with event ordering, idempotency, callbacks, and exception flows. It thus becomes even more important to architect systems the right way the first time and not fall into the trap of just following the trend.
2. Don’t run around in circles - Use In-Memory data grids.
Database trips come with expensive tickets. Often, this is a choking point for many applications. It is always a good practice to refer to read-only and master data from a nearby cached data store. While there are many efficient caching solutions available, choose one that has in-built persistence or rehydration capability, is cloud-ready, supports replication, and is scalable.
3. As quick as a flash - Process messages with speed & efficiency.
The ISO 20022 messaging protocol is now a de facto standard for the financial industry, considering the limitations in legacy message protocols for new use cases in real-time and cross-border payments. Global real-Time payment platforms widely use the ISO 20022 standard or their proprietary message formats. In either case, XML is heavily used. Parsing XML messages is CPU- and memory-intensive. Hence, it is important to ensure that your system deploys the most efficient parser, ensures the message is parsed at the right stage, and parses only what is needed.
While choosing a serialization format, a balanced decision should be made considering speed and complexities in design and development. Also, the usage of internal data structures to efficiently transform, retrieve, and relay required data between payment services yields remarkable results.
4. Keep a close watch – Ensure code instrumentation & observability
The ability to identify performance bottlenecks proactively is a key architectural attribute for any system. Each component in the system architecture should be observable to prevent failures and to instantly recover from failures. Most of the widely used technology stacks that work within the purview of the EDA framework have the capability to publish performance and usage metrics that can be exported to industry-standard visualization tools. Similarly, code instrumentation tools in microservices frameworks provide useful performance metrics that can be plugged into any visualization tool. Moreover, when using EDA and MSA, key measures such as rate of processing and queue sizes offer invaluable insight on the system’s health.
5. Look before you leap - Pick your technology platform wisely.
When choosing a technology platform to implement an architectural style, remember that there will be no turning back at advanced stages.
The following advice can be helpful throughout the selection process:
While these are just a few application design aspects, it is extremely important to take a holistic view of the solution, including DevOps practices, Deployment Topology, Security Architecture, Database Partitioning/Sharding techniques, and Infrastructure components.
The Silver Bullet
Financial institutions can stay ahead of the curve in the ever-evolving payment landscape by choosing to modernize their payment system architecture for real-time payments and other innovative use cases. Consolidating all payment processing requirements into one and incremental modernization are a few approaches that can be considered for this initiative.
To align with such architectural ideologies, REN was purpose-built from scratch to emerge as a reliable solution for Instant Payment System Processors/Operators and Participating FIs.
Reminiscing Futurist Jim Carroll’s words to conclude - “The future is inevitable. Your decisions and actions are not!”. Get in touch with the Ren team to know how Ren can solve your real-time payment processing needs and watch this space for more on payments.
The rapid evolution of payment technology demands forward-thinking strategies. Our guide offers detailed insights to help you modernize your systems and keep pace with industry advancements. Prepare for the future—download your guide today!
Comments