Introduction
The solution architecture was designed and implemented by Hutech bottlenecks of UPI transactions for one of the top private sector banks of India, generally placed around 3rd to 4th by market capitalization.
India is the first country to implement digital payments at the population scale, speed, and public infrastructure level.
UPI (Unified Payments Interface) is India’s real-time digital payment system that allows people to send and receive money immediately using their smartphones, launched in 2016. Features such as real-time transfer, 24×7, using a simple identifier called UPI ID, phone number, QR code issued by any bank to the account holder, secured transaction between bank accounts using apps like Google Pay, PhonePe, Paytm, BHIM, and many bank apps, make it special and popular.
This was by the National Payments Corporation of India (NPCI), officially inaugurated by the Reserve Bank of India along with major Indian banks. The NPCI is owned by a consortium of Indian banks, including SBI, HDFC, ICICI, Axis, and many others.
UPI handles 640–700+ million daily transactions. In 2024, it processed 172 billion transactions worth ₹ 246.8 trillion. Monthly volumes surged past 20 billion transactions from October 2025, and NPCI estimates the 2025 transaction value at ₹300 trillion, with volumes expected to top over 200 billion transactions.
In the NPCI ranking of UPI beneficiary members, our customer bank features among the top banks with significant monthly volumes of 350 + million monthly transactions.
Problem Statement
The rapid expansion of UPI transactions introduced unpredictable spikes in traffic and heavy workloads of inquiries that could not be efficiently absorbed by the legacy core banking systems. The bank set up a technical team and engaged our engineers and architects for the design, implementation, and testing of a scalable AWS architecture that protects the core systems while delivering reliable and low-latency performance.
Challenges
With many more inquiry requests, including read-only requests, during periods of peak transactions, high loads were placed on the core banking system, which caused performance bottlenecks. This increased pressure introduced latency and availability risks, impacting the system’s capability to deliver consistently reliable customer experiences.
Architecture Highlights
The solution architecture consisted of a cloud-native AWS architecture, which enhanced overall transaction performance related to UPI and mobile banking by enhancing the performance of the existing core banking platform.
The cloud-native AWS architecture that segregates real-time transaction processing from high-volume workloads. This decoupling facilitated horizontal scalability and operational resiliency to ensure predictable, low-latency performance of UPI payments and mobile banking transactions during peak periods of use.
The cloud-native AWS architecture uses Amazon Elastic Kubernetes Service (Amazon EKS) for scalable microservices orchestration, Amazon Managed Streaming for Kafka (Amazon MSK) for near-real-time data streaming and synchronization across systems, and Amazon ElastiCache for Redis to achieve high-performance in-memory caching for fast response times.
This architecture greatly enhanced the reliability of processing UPI transactions, inquiries into balances, transaction history, and mobile banking, enabling uninterrupted customer access even under partial degradation of the core banking system (CBS).
Key scalability and performance challenges were addressed in consultation with the bank through the design and delivery of two complementary systems: RTS and ROS.
RTS serves as a distributed microledger, tracking all value movements in real time, while ROS offloads high-volume, non-transactional workloads from the core banking system.
Using Amazon DynamoDB at its core, this approach provides predictable low latency and horizontal scalability that enables the bank to serve the demands of modern digital banking and high-throughput UPI and mobile transactions.
Amazon Elastic Kubernetes Service (Amazon EKS) is used to orchestrate RTS and ROS microservices, while Amazon DynamoDB acts as a highly scalable and resilient micro-ledger. An event-driven model is enabled through Amazon Managed Streaming for Kafka (Amazon MSK) to support near-real-time data propagation and system decoupling. Transactional and analytical data is stored in Amazon S3 and analysed with the use of Amazon Athena, supporting reporting and insights, complementing the bank’s Finacle core banking database.
To continue alleviating the burden on the core banking system, a dedicated Account Inquiry Service, or AIS, was developed. AIS caters to inquiries about balances and transactions using a separate, read-only representation of account data stored outside the core banking platform. Because inquiry workloads are separated from transactional processing, AIS significantly reduces response times while it provides a very stable system and continuous customer access to their account information during peak load situations or partial degradation of the CBS.
Outcomes
The implemented solution provides consistently low-latency digital transactions while considerably reducing the load on the core banking system. Architecture enables the platform to scale efficiently and handle sustained growth in UPI transaction volumes by fostering overall system resiliency.
Hutech's Role
Our work included the implementation of architecture design on AWS, integration with existing banking systems, testing, and providing rollout and deployment support to ensure a smooth transition into production.
Frequently Asked Questions
The bank faced performance bottlenecks due to massive UPI transaction volumes and high inquiry workloads that stressed legacy core banking systems during peak traffic.
Hutech implemented a cloud-native AWS architecture that decoupled transactional processing from inquiry workloads, introduced microservices orchestration, and added scalable caching and streaming systems.
RTS is a distributed micro-ledger handling real-time value movement tracking. ROS offloads high-volume, non-transactional requests from the core banking system.
The Account Inquiry Service (AIS) separates read-only queries from transactional systems, reducing load on the core banking platform and ensuring consistent response times.
MAIL US AT
sales@hutechsolutions.com
CONTACT NUMBER
+91 90351 80487
CHAT VIA WHATSAPP
+91 90351 80487
Humantech Solutions India Pvt. Ltd 163, 1st Floor, 9th Main Rd, Sector 6, HSR Layout, Bengaluru, Karnataka 560102
Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ut elit tellus, luctus nec ullamcorper mattis, pulvinar dapibus leo.
