Hello to Amasty blog readers. We continue to publish articles on using cache for Magento.
Our customers frequently ask us how much faster Amasty Full Page Cache is, compared to Magento Enterprise Full Page Cache. And we couldn’t answer this question with precise numbers, as we didn’t actually run comparison tests. As the number of questions grew with time, we finally decided to run a test and show some exact results to our customers.
Also, we are often asked a question: why do all the extensions on our demo server work faster than those installed on our clients’ servers?
Such a result is obtained thanks to what is under the hood of our server: it’s a dedicated server with powerful processors, large RAM size and a fast SSD storage.
So, in order not to confuse our readers and to push the results closer to reality, we used a virtual server with more common configuration. Also, this configuration is what we always advise as minimal required setup for working with Magento.
Server configuration for FPC comparison test
CPU: Intel Core i5-4570 CPU @ 3.20GHz
Storage: SSD 20G
OS: Debian GNU/Linux 7.8 (Wheezy) amd64
Magento stack: Nginx 1.2.1, PHP-fpm 5.4.39, MySQL 5.5.43, Magento Enterprise Edition 184.108.40.206, Sample data 220.127.116.11
FPC comparison test procedure
As always, we use Apache jMeter for performance tests. I ran rounds of tests for 10, 20, 30, 40, and 50 simultaneous users. We have already described the test procedure in the previous Amasty research: Full Page Cache vs Varnish Cache article and were asked to share the test so the users could run it on their own servers.
Today we’re sharing an updated scenario with sample data 18.104.22.168 support, feel free to download it here.
When testing Amasty Magento Full Page Cache, we turned off the built-in Enterprise Full Page Cache by editing app/etc/modules/Enterprise_PageCache.xml file to make sure that Enterprise FPC code won’t influence the test results.
Similar to this, Amasty Full Page Cache was turned off by editing app/etc/modules/Amasty_Fpc.xml and app/etc/modules/Amasty_Fpccrawler.xml files when testing Enterprise FPC.
Here’s the procedure scenario:
- go to homepage
- search ‘shirt’
- filter by ‘sale’ category
- go to ‘Slim fit Dobby Oxford Shirt’ product page
- add the product to cart
- go to ‘Men / Blazers’ category
- go to ‘Men / New arrivals’ category
- add ‘Linen Blazer’ product to comparison
- add ‘Khaki Bowery Chino Pants’ product to comparison
- show product comparison page
Thus, the procedure scenario consists of 10 actions, performed in parallel for 10, 20, 30 etc users, and each test iteration is repeated 10 times, so we can calculate an average time of each action execution and an average time of the whole iteration as well. If an error occurs during the test execution, the execution stops. By the way, we didn’t get any errors during the procedure.
Here are the test results:
To give you better understanding of the results, we built some charts by number of users:
Amasty FPC vs Magento Enterprise FPC test results
As we can see from the charts, Amasty Magento Full Page Cache is ahead of Magento Enterprise Full Page Cache for most of the actions.
To give you the understanding of how this works for your business, here’s the table of performance improvement you can get if you replace Magento Enterprise Full Page Cache with Amasty Magento FPC.
In short, Amasty FPC adds around 20% to your Magento store’s performance, though the results are different depending on the action, worse for comparison and much better, up to 40% performance growth for some points.
I hope this article was helpful for you. If you have any questions about the test or the work of Magento Full Page Cache, I’ll be happy to help.