Using Memcached to Speed Up Data-Heavy MODX Projects

When a project starts to outgrow file-based caching, the next useful step is often not a redesign. It is better cache storage. This older article introduced Memcached as a way to keep associative data in memory and reduce repeated work inside PHP.

Basic installation

Ubuntu: sudo apt-get install memcached
Gentoo: sudo emerge install memcached
RedHat: sudo yum install memcached
sudo pecl install memcache

Once installed, the service can be started with a dedicated memory limit, IP, and port:

memcached -d -m 1024 -l 10.0.0.1 -p 11211

Using it from PHP

The original note extended an xPDO-based class and connected the cache manager to Memcached directly:

class MyXPDO extends xPDO
{
    function __construct($dsn, $username = '', $password = '', $tablePrefix = '', $driverOptions = null)
    {
        parent::__construct($dsn, $username, $password, $tablePrefix, $driverOptions);
        $this->config = array_merge(array('cache_db_handler' => 'Memcache'), $this->config);
        if ($cacheManager = $this->getCacheManager()) {
            $cacheManager->objcache->connect('10.0.0.1', 11211);
        }
    }
}

Why this mattered

On busier projects the gain was not theoretical. Cached objects, prebuilt blocks, and repeated query results all became cheaper to serve. The original article also pointed out the next real-world question: object caching is useful, but busy news or article projects also need a strategy for page fragments and invalidation.

That is still the important lesson. Better caching is not just about speed. It is about deciding what to cache, where to cache it, and when to refresh it.