For test purposes, Broadleaf focuses on throughput as it relates to completed transactions per second (TPS), and completed orders per second (OPS). By adding replicas of application microservices to a growing Kubernetes cluster, we demonstrate fairly linear scale as we seek to increase throughput.

In this round of tests, the Balanced FlexPackage™ throughput was examined through small, medium, and large infrastructure configurations, culminating in 3400 TPS for the system. The Granular FlexPackage™ was also considered at the larger infrastructure sizes and performed comparably.

Data Variations

Discounts can slow down cart manipulation as it introduces additional calculation required to price items in the cart. Variants (aka SKUs) and product options also represent additional catalog complexity for the system to navigate as it addresses inventory and pricing concerns across a larger landscape of customer choice.

100 BOGO offers were introduced into the system - all set to auto-apply. This setting indicates that all offers are qualified against the cart contents at the time of add to cart, which is a key computational moment in the cart lifecycle. 3 variants per product were introduced with a basic set of product options for the test to negotiate upon add to cart. When compared to normal execution (no variants), the impact was negligible, within a margin of error.

Bar Graph

Catalog Size

For corporations requiring larger catalog sets, Broadleaf tested an online catalog with 100,000 and 1,000,000 products. This test focuses on the order life cycle journey using a test plan starting with the product detail page and ending in cart checkout.

Again, we see a negligible impact to performance based on catalog size across all infrastructure combinations.

Conversion Rates

For corporations with increased conversion rates above industry averages, Broadleaf tested up to 100% conversion. The key metric captured for this test is orders per second (OPS).


Concurrent Users

In order to simulate concurrent users in a real-world scenario, Broadleaf tested virtual users in gated test plans involving percentage advancement at various stages of the customer journey, coupled with quantity and catalog choice randomization. Then, calculating based on an eCommerce average of 1 page view per minute per user, we were able to estimate concurrent user capacities.
For most businesses with typical conversion rates, the numbers demonstrated above can handle sites generating billions of dollars in sales.
Concurrent Users

Broadleaf Commerce: Scalability Study

Want to see the proof? Fill out the form to download the Broadleaf Scalability Study!

Must fill in captcha below *

Your request has been received

All requests are promptly responded to via email.

Please check your inbox.

Your Download is Ready!

Downloading file…

Broadleaf Scalability White Paper-3.pdf

File did not download correctly? Click here to manually download file.