Open Source Projects:
Pricing
Chat
Contact
Back to top

High-Speed Transaction Processing with No Compromise on Fraud Detection

October 01, 2019

Transaction processing in today’s global payments networks is what essentially powers the global economy. The framework for transaction processing has expanded and morphed considerably over the past few years as consumers have started using their mobile devices to execute payments. Meanwhile, non-traditional deep-pocketed players like Google and Apple have entered the system with competitive alternatives, not to mention the fact that everyone everywhere is online all the time, and more often than not, they’re spending money.

The implications for payment processors, banks, card providers, and others in the supply chain is that the volume of transactions has skyrocketed. When we say volume, it means that at leading payment processors the transaction rates are measured in millions of events per second.

There are two things to consider here; first, “transaction” can imply something fairly complex, which for example can mean purchasing a mobile device online which requires a complex configuration, shipping and tax variables, add-ons like premium support, etc. When a major device provider announces the latest offer, millions of people hit the site at the same time, and all those transactions need to be processed in real-time, with no risk of fraud.

The second factor is speed. When you’re at the checkout stage of a transaction, how long does it take before you start contemplating visiting a different retailer? Probably not more than a few seconds. In that time, the merchant needs to pull together a complex set of purchase parameters and get verification from the bank card provider, who then approves or declines the purchase. During this period, the transaction needs to traverse a network, reach the payment processor’s gateway, where the card provider needs to verify that the card is valid, run through a series of fraud detection algorithms at sub-millisecond speed, get approved (or denied), then traverse back across the network to the consumer with an accept or decline message. All of this happens in about two to three seconds, and most of it happens in less time than it takes to swipe a card. And just to make it interesting, millions of people are doing this at the exact same time.

So in this world, taking the time to access relevant information in a database (which can take multiple seconds or even minutes) is completely out of the question. Running multiple fraud detection algorithms at sub-millisecond speeds is not what a database was designed to do; their primary function is storage, not processing. Fortunately, in-memory technologies can deliver the speed required to successfully participate in the digital payment ecosystem. It starts with implementing a stream processing engine (an application specifically designed to process high volumes of data at real-time speeds) that are coupled with an in-memory data grid (which keeps information in a cache that’s accessible at RAM speeds). Now, the idea of processing millions of transactions per second and running multiple fraud detection algorithms in under 50 milliseconds (for reference, you blink in 300 milliseconds) suddenly is not only feasible, but it also becomes a low hurdle to cross.

This is already happening with some of the biggest e-commerce and payment processing companies in the world who also happen to be Hazelcast customers. Millions of transactions per second? Check. Incredibly robust fraud detection? Check. Elastic scalability? Check. Rock-solid stability? Check. To get a better understanding of how in-memory technology can take your game to the next level, dive into our resources page.

About the Author

About the Author

Dan Ortega

Dan Ortega

Product Marketing

Dan has had more than 20 years of experience helping customers understand the business value of technologies. His domain expertise spans enterprise software, IoT, ITSM/ITOM, data analytics, mobility, business intelligence, SaaS, content management, predictive analytics, and information lifecycle management. Throughout his career, Dan has worked with companies ranging in size from start-up to Fortune 500 and enjoys sharing insights on business value creation through his contributions to the Hazelcast blog. Dan was born in New York, grew up in Mexico City, and returned to get his B.A. in Economics from the University of Michigan.

Latest Blogs

Right-Sizing Your On-Demand Infrastructure for Payment Processing in Any Cloud

Open Banking and the Application of In-Memory Technologies

Open Banking and the Application of In-Memory Technologies

The Role of Streaming Technology in Retail Banking

The Role of Streaming Technology in Retail Banking

View all blogs by the author

Free Hazelcast Online Training Center

Whether you're interested in learning the basics of in-memory systems, or you're looking for advanced, real-world production examples and best practices, we've got you covered.