This post is part of a series: FirstPreviousNext

Introduction

We’re past the halfway mark for our e-commerce benchmarking experiments where we score a bunch of open source e-commerce platforms for performance. Next up is VirtueMart, which is an e-commerce extension for the Joomla content management system.

The first time through, I had anticipated VirtueMart to be a standalone e-commerce platform. VirtueMart’s official site also doesn’t mention Joomla on the front page, or that VirtueMart is an extension. It’s funny because VirtueMart’s official Twitter page shows that information but their official website doesn’t.

The VirtueMart Sample Data Store

Out of all the web stores I had encountered, VirtueMart was the most difficult to set up. Many times I had to reinstall not only VirtueMart, but Joomla too. I kept receiving the dreaded PHP “white screens of death” when visiting product pages. This was partly attributed to the missing the php-xml package.

After resolving that issue, I was greeted by a store with dead links! I had to manually fix those in the VirtueMart backend.

If you decide to use VirtueMart, I would recommend hiring someone who knows what they’re doing. A graphic designer is not a bad idea either because out-of-the-box VirtueMart is lacking aesthetically. After all, it’s just an extension.

Testing Phase

Using the LoadStorm load testing tool, we executed the testing phase on VirtueMart to benchmark it with the other e-commerce platforms. The following charts show our load test results:


VirtueMart Three Load Test Results (click to expand)

The test results are pretty consistent, but will that mean a high score for VirtueMart? We estimate a sustainable scalability up to 1012 concurrent VUsers based on our criteria.

We see performance errors take a sharp increase at 10 minutes. Most are Request Connection timeout errors, meaning some VUsers are having trouble even establishing an HTTP connection to VirtueMart.

Other problems point to the almost immediate jump in peak response time of 15 seconds near the beginning of the test. All it takes is one slow resource, and you get extraordinarily long times.

What will bring down VirtueMart’s score significantly compared to other platforms is its average response time. Unlike peak response time, average response time should definitely be kept low at all times. With an average response time of 10.06 seconds, that’s going to seriously affect VirtueMart’s score.

Performance Errors for VirtueMart Test #3

Before we call it quits, we still have to factor in the WebPageTest results into our scoresheet for a conclusive VirtueMart benchmark. Maybe those results will save this e-commerce platform from a low score. Unfortunately, the WebPageTest load test averages don’t help much.

Scoresheet


It appears that VirtueMart has also scored lowest out of all the e-commerce platforms tested so far. With a score of 54.28, it’s not looking good. There are still two remaining e-commerce platforms remaining to test.

I’ve never tried Opencart or Drupal Commerce so all bets are still up in the air.

This post is part of a series: FirstPreviousNext

Similar Posts