Follow Datanami:
September 11, 2019

Hazelcast Enhances Real-Time Capabilities for Financial Services Industry

Palo Alto, Calif., Sept. 10, 2019 – Hazelcast, the leading in-memory computing platform that delivers the System of Now, continues to see rapid adoption among the world’s largest financial institutions across dozens of use cases, including fraud detection and payment processing. Hazelcast provides the top 50 financial institutions with ultra-high performance for time-sensitive, cloud-native applications. Among its current customers, Hazelcast accelerates business-critical applications at JPMorgan Chase, Lloyds Banking Group, UBS, National Australia Bank and many more.

To help financial services organizations deliver a better, real-time customer experience, the Hazelcast in-memory computing platform uniquely integrates proven data grid technology with next-generation event stream processing capabilities. Hazelcast directly addresses the inherent limitations of databases by providing ultra-low latency at an extreme scale for stored data while running the industry’s fastest streaming engine with run-anywhere capabilities. The extremely lightweight footprint allows financial services organizations to accelerate their business applications on any cloud or edge device, in addition to on-premises data centers.

“Given the data-centricity and time-sensitive nature of financial services, advanced in-memory technologies are increasingly being deployed in new use cases where time is money and latency is the enemy,” said Kelly Herrell, CEO of Hazelcast. “With its unique ability to process both stored and streaming data at next-generation speed, the Hazelcast System of Now can radically accelerate the performance of not only existing architectures, but the rapidly-growing cloud architectures and applications leveraging ML/AI for a competitive advantage.”

Fighting Fraud In-Memory

Three of the four top global card processors are among Hazelcast’s customers in the financial services sector. Digital and mobile card payments have dramatically increased both transaction volumes and attack vectors for fraudulent activity, taxing traditional infrastructure for fraud detection. As evidence of the increasing and evolving threat landscape, a survey1 of 5,500 consumers identified shopping and banking as the top uses for smartphones, further increasing the volume of transactions. However, the service level agreements (SLAs) agreed upon by vendors and payment processors have remained unchanged or shrunk.

Once a transaction request arrives at a credit card processor’s data center, they typically have only milliseconds to match account information, check available balances, execute fraud detection and render a decision. For example, until recently, a tier 1 US credit card processor was writing off $1.5 billion a year due to missed fraudulent charges. With only 50 milliseconds to execute its detection algorithms, the processor was running into the transaction rate limit of its legacy relational database architecture and reduced the number of employed detection algorithms. By moving to an in-memory architecture with Hazelcast, the company is able to reallocate the time normally spent accessing an external database and deploy numerous additional algorithms. This architectural change helped the company save hundreds of millions of dollars in fraudulent charges while ensuring that they could scale their fraud infrastructure to meet the needs of today’s digital and mobile users, even during peak events such as Black Friday and Cyber Monday.

Achieving Zero-Downtime for Real-Time Payment Processing

Whether it’s real-time payments, P2P payments, open banking, ISO 20022 or other compliance-driven requirements, the window for processing and executing financial transactions is shrinking, while the expectations for zero downtime have increased to ensure 24 x 7 x 365 access. Furthermore, any delays in payments not only draw scrutiny from merchants, businesses, and consumers but from government regulators as well. An in-memory computing platform with integrated stream processing enables faster, more scalable and more resilient payment applications.

In the event of a network or disk failure in a legacy database architecture, the data is inaccessible until the system recovers and the backlog of transactions are processed. In-memory computing solutions are distributed by design and Hazelcast offers additional critical capabilities to ensure extraordinarily high system availability. Additionally, in combining Hazelcast’s multi-cloud, run-anywhere approach with features such as Hazelcast WAN replication, financial organizations can achieve new levels of data resiliency by spanning across data centers and regions to deliver zero-downtime applications.

Recent in-memory technology advancements by Hazelcast enable financial service organizations to adopt a consistency or availability strategy in the event of a network failure. Under the consistency approach, a system will only deliver information if it is the latest, whereas the availability approach is guaranteed to respond regardless if it is accurate. These capabilities ensure data integrity and are currently only available from the Hazelcast in-memory computing platform.

[1] Experian, “The 2018 Global Fraud and Identity Report” https://www.experian.com/assets/decision-analytics/reports/global-fraud-report-2018.pdf

About Hazelcast, Inc.

Hazelcast delivers the System of Now, an in-memory computing platform that empowers Global 2000 enterprises to deliver innovative, low-latency, data-centric applications. Built for ultra-fast processing at extreme scale, Hazelcast’s cloud-native in-memory data grid and event stream processing technologies are trusted by leading companies such as JPMorgan Chase, Charter Communications, Ellie Mae, UBS and National Australia Bank to accelerate business-critical applications.


Source: Hazelcast

Datanami