TEK Webcast Notes

« »

Last Friday, I did a webcast as a part of the PHP TEK Webcast series. The webcast was on Caching for WordPress. We had a good turnout, and there were lots of questions about the best plugins to use for WordPress caching.

For me, every WordPress blog should have APC installed on it, hands down. APC almost doubles the response rate for WordPress by itself; APC is easy to install and freely available, stable and exceptional. Along with APC I also recommend either WP-Super-Cache or W3 Total Cache. WP-Super-Cache is great for sites that simply want to do static caching; W3 Total Cache is for sites that want to make use of S3 or other caches (like the APC User Cache).

There were also questions asked about my benchmarks of each of the caches. In short, I used Apache Bench to get my benchmarks. The benchmarks were conducted from a development server on Slicehost, with a 1 GB slice. The web server and the test server were on the same machine. I used 3,000 requests (no concurrent requests) per benchmark, with the exception of WordPress by itself (which used 1,000 requests, because it was so slow). All the caches were configured properly according to how the author recommended. I did test to make sure the components were behaving as advertised before benchmarking them.

You can download the slides here. Thanks to everyone who came out!

Update: Cal Evans has posted the video of the webcast here.

Brandon Savage is the author of Mastering Object Oriented PHP and Practical Design Patterns in PHP

Posted on 4/21/2010 at 7:00 am
Categories: Conferences, Best Practices, Technology
Tags: , , , , ,

Shaun Farrell (@farrelley) wrote at 4/21/2010 7:44 am:

Excellent presentation. We are actually in the process of implementing WP-Super-Cache and some other caching alternatives on our systems int he coming months.

Clarification – On your graphs what do the different color bars represent? Red, Gray, and Brown?

Brandon Savage (@brandonsavage) wrote at 4/21/2010 7:48 am:

The colors represent each run of the benchmark. I did each benchmark 3 times to avoid outliers, errors, etc. So all three tests for all the benchmarks are shown on the graph.

Gary (@garyj) wrote at 4/22/2010 2:38 pm:

It’s also worth pointing out that the CDN aspect of W3 Total Cache can also be used from, say, a subdomain of the main site.

You won’t get the speed of S3, and the data transfer costs will still be your own, but you’ll benefit from parallel domain downloads.

It’s a halfway house between the default, and a full S3 CDN.

Thanks for the presentation!

« »

Copyright © 2024 by Brandon Savage. All rights reserved.