Understanding Throughput in Financial Transactions

Published on
Jun 21, 2024
Written by
Read time

Throughput is an interesting concept. While it seems straightforward, like the rate of water flowing through a garden hose, it becomes more complex when dealing with financial transactions. Not all transactions are created equal.

In financial transactions involving buying or selling, there's usually an invoice. This could be a sales order, a purchase order, or items from an online shopping cart. Depending on the type of transaction, this “invoice” may include various details such as purchase location, shipping information, product or service line items, quantities, discounts, and more.

From a sales tax engine perspective, all these invoice data points must be collected, evaluated, and processed for accurate tax calculation. Each invoice, or transaction, can vary significantly in the data it contains. For instance, one transaction might be a sales invoice with hundreds of product line items, while another might be transactions from an e-commerce site as items are added to a shopping cart. Other transactions could involve real-time cell phone data from a telecom provider or multiple telecom Voice-over-IP switches sending requests simultaneously.

Processing these transactions for sales tax calculation can be challenging. It takes time to determine what data is contained, process it, and deliver the correct tax amounts. When the rate of inbound data to the tax engine is very high, originating from numerous and diverse sources, most systems struggle to handle the throughput. This challenge is compounded when multiple high-volume events occur simultaneously, such as processing Black Friday transactions while others are submitting large batches of invoices from an accounting system. We have historically seen existing sales tax engines fail under these conditions—they simply cannot handle the data burst.

For the production release of CereTax, we focused extensively on designing ways to handle data bursts. We support financial transactions or invoices from many different vertical markets, each with its own complexities. At launch, we successfully processed 10,000 simultaneous transactions per second for a prolonged period with no errors. This level of processing throughput is not affected by the type of invoice or whether it contains one line item or 500. However, this rate is not our limit; we considered it a fair representation of a burst event that would overwhelm most sales tax engines.

One advantage of delivering a modern sales tax engine is the ability to use current approaches for system architecture and data management. We are fortunate to incorporate these techniques into the CereTax design. Our clients can send transactions to our tax engine at any burst level—they can trust that we understand throughput and will not break.

Click here to connect with a team member to learn more about CereTax can simplify financial operations for you business.

Mike Sanders Mike Sanders
Co-Founder and CEO, CereTax