Zafin streamlines operations and shortens time to value for world-leading banks

Challenge

Finding a simpler, high-throughput Kafka alternative  

Zafin is a SaaS that drives core modernization and transformation, empowering the world’s top banks to stay ahead of the competition in the digital era. Zafin was running on Confluent, had 80 Apache Kafka^® brokers in two regions, and was processing up to 800 millions of events per day.

However, Zafin’s engineers were burdened with Kafka’s high maintenance and struggled with frequent outages. One major outage took out both regions at once.

“Banks aren’t willing to modernize if they risk downtime, poor security, or compliance issues,” explains Shahir Daya, CTO at Zafin. So, to continue its mission of increasing everything from revenue to customer experiences for banks, Zafin needed a streaming data architecture that was simple to set up, easy to manage, and that “just worked.”

Why Redpanda

Simplified streaming data and simpler operations

While researching Kafka alternatives, Shahir read about a streaming data platform built in C++ and was hooked. After asking the team to run a proof of technology, Zafin readily chose Redpanda due to its:

• Simple day 2 operations
• Smooth developer experience
• Full compatibility with Apache Kafka^® APIs
• Freedom from JVM and external dependencies
• Responsive support team

During their proof of technology, the Zafin team replaced Confluent with Redpanda and was shocked to find everything was still running. “That was the game changer,” Shair says. “We didn’t need to change our code, and suddenly, our day two operations were much simpler.”

Results

A reliable system of reference for real-time events

The Zafin Integrate & Orchestrate (Zafin IO) platform integrates real-time data between a bank’s core systems and Zafin’s. One of its first adopters was a world-leading bank that was producing more than 300 million daily events and relied on Zafin to efficiently ingest and process that real-time data to surface business insights.

With Redpanda at the center of ZafinIO, stream processing applications ingested the bank’s real-time events into Redpanda and transformed them into a unified industry-aligned data model. Alongside Apache Flink, Apache NiFi, and Temporal, this streamlined data pipeline allowed Zafin to centralize the bank’s data for comprehensive insights, eliminated the need to integrate with multiple systems of record, and shortened the time to value. 

Furthermore,  the new unified system lowered compute costs and simplified day 2 operations for Zafin. Gone were the days of complex maintenance and frequent outages. With Redpanda at its core, Zafin can provide banks with a reliable system of reference for real-time events, seamlessly processing up to 800 million events every day.

“With a single platform that makes life easier for our operations team, Zafin is now dedicated to going all-in with Redpanda,” says Shahir. So before the end of 2024, Zafin will replace five brokers and five ZooKeepers for every client with just three brokers of Redpanda.

Shahir also credits the move to Redpanda’s reliable guidance and troubleshooting. “Redpanda has a fantastic team that is exemplary of how to provide support,” he says. “It gives us the confidence to embark on modernizing some of the biggest banks in the world.” 

Replacing their Kafka implementation with Redpanda was only the beginning. As ZafinIO continues to get adopted by more banks, Redpanda’s footprint will grow along with it.