Real-Time Transformation: How a Top 3 U.S. Bank Modernized Its Data Analytics with StarTree

The role of Global Spread Products

The Global Spread Products (GSP) Group at a top 3 US bank provides essential capital solutions to clients by offering liquidity and innovative financing across a wide range of credit and securitized markets. The group’s activities include origination, structuring, investing, lending, and market making across products such as corporate and emerging market bonds, asset-backed and mortgage-backed securities, collateralized loan obligations, short-term interest rate products, loans, letters of credit, and derivative instruments.

In addition to project financing, securitization, and warehouse lending, the team plays a critical role in underwriting, secondary trading, and commercial, residential, and affordable housing financing. These operations require complex financial decision-making to ensure optimal capital allocation and risk management in real-time markets.

Given the broad range of activities and instruments, the GSP team deals with massive volumes of fast-moving data that must be processed and acted upon immediately to reduce risk and maximize returns. In this environment, slightly stale data can lead to significant financial exposure, suboptimal pricing, and missed opportunities in fast-moving markets. The need for real-time insights across multiple data dimensions is critical to maintaining the bank’s competitive edge.

The challenge: Complexity and stale data were slowing decisions

Prior to adopting StarTree, the company’s GSP team relied on a Kafka-Flink-KX (a time-series database developed in the early ‘90s and still widely used in FinServ) architecture to process and deliver data insights.

While this system enabled real-time data ingestion and analytics, it had several limitations that impacted both the business and IT teams:

• 10-minute data freshness: The architecture resulted in up to 10 minutes of data latency, which could expose the bank to financial losses due to outdated insights in fast-moving markets.
• Difficult to scale: The system required multiple KX instances to manage different use cases, leading to high infrastructure costs and complex operational overhead. KX uses a highly specialized, non-standard programming language (q), which limits broader adoption and inflates people costs due to the scarcity of developers skilled in the language, making it harder to scale and maintain.
• Restricted set of users: Due to scalability constraints, only a limited number of users could access real-time insights, creating bottlenecks in decision-making processes.
• Slow time-to-market for new data sets: Incorporating new data sets into the analytics platform was time-consuming and inflexible, which reduced the team’s agility to adapt to changing market conditions.

This legacy system limited the bank’s ability to act on real-time data insights, increasing their exposure to financial risk and reducing capital utilization efficiency.

The solution: Modernizing financial analytics with Apache Kafka and StarTree

To address these challenges, the bank adopted a new architecture based on Kafka and StarTree. By replacing Flink and KX with StarTree’s real-time analytics platform, they were able to achieve faster, more scalable, and more flexible data insights.

The new architecture offers significant improvements in both data freshness and complexity reduction:

• Kafka for real-time data ingestion: The GSP team continues to use Kafka for real-time streaming, ensuring continuous data flow into StarTree.
• StarTree for real-time financial analytics: StarTree provides subsecond query performance on high-concurrency workloads, enabling fast, interactive insights across complex, multi-dimensional data sets.
• Trino for query layer: Trino is used as a flexible query layer and data fabric to enable ad-hoc querying and seamless data access across various sources.

The results: Business and IT benefits with StarTree

Business outcomes

• 10-second data freshness: GSP team reduced data latency from 10 minutes to 10 seconds, allowing the GSP team to make faster, more informed decisions in high-risk markets. This reduction in latency significantly lowers the risk of financial loss and improves capital utilization.
• More dimensions to inform actions: StarTree enables the GSP team to query more data dimensions in real time, providing richer insights that support more precise pricing, risk assessment, and market-making decisions.
• Expanded set of users: The new architecture scales easily, allowing more users across the bank to access real-time insights without bottlenecks, improving collaboration and decision-making across the organization.
• Faster time-to-market for new data sets: StarTree’s agility allows the bank to quickly onboard new data sets and adapt to changing market conditions, providing a competitive advantage in fast-moving financial markets.

IT outcomes

• Simplified architecture: By consolidating to a Kafka-StarTree architecture, this bank significantly reduced the complexity of its data stack, eliminating multiple KX instances and reducing the number of hops in the data flow.
• Easier scalability: StarTree’s architecture is designed to scale with high-concurrency workloads, reducing operational overhead and making it easier to manage as data volumes grow.

Conclusion

In financial markets, every second counts. With StarTree, large banks can now make real-time, data-driven decisions with confidence and agility, reducing their exposure to market risks and improving capital utilization. By simplifying their architecture and improving data freshness, this bank has transformed their financial analytics capabilities, positioning themselves for continued success in the fast-evolving financial landscape.

Get started today

You can get a free developer account on StarTree Cloud. There’s no trial limit to it. Use it for free, forever. When you’re ready to move into production, you can move into one of our cost-effective packages just right for your business.