By now, WordPress has a worldwide market share of a good 65 percent among content management systems (CMS). Nevertheless, WordPress is sometimes still ridiculed as a small CMS. But WordPress does not have to hide at all - especially not when it comes to high performance. What WordPress can do in this field, we'll show you today.
How your magazine survives 75,000 requests per minute
Difference between calls and visits
Before revealing the secret of how your website can handle extremely high numbers of visits, I'd first like to explain the relationship between visits and views. One person can, of course, visit several subpages. The reflection time between clicks is the key indicator to establish the relation between visits and views.
If 900 people click on a website every three seconds on average, we have 300 views per second for 900 visits.
As an important rule of thumb, the number of visits is generally at least twice as high as the number of views.
WordPress & High Performance - Caching as a Magic Bullet
Our caching is not a separate Varnish server, which is inconveniently connected upstream, but a technology that is integrated by default on the server side.
Thousands of calls and the Processor gets bored
Due to the server-side caching, the requests are delivered directly without even contacting the processor and the database. In our starter tariff, for example, 75,000 calls can be answered within one minute by default. Especially for visitor-intensive websites such as blogs, magazines and those of larger companies, WordPress is therefore an excellent system.
This is how we often experienced it during high performance times, for example after certain websites were mentioned in a TV show or during online marketing campaigns - even though the server was upgraded in advance, the processor load usually remained at a low level because the caching had largely taken over the delivery of the pages.
High Performance Special Case - WooCommerce
However, there is one important limitation: The statement refers to static sites , which can be stored in the cache.
Dynamic requests can poorly be cached
For dynamic requests, such as filling out contact forms or checkout processes in the shopping cart, caching is not allowed.
An example of dynamic content are product pages where the prices in the shopping cart change depending on the interaction with the website. Here, it would be fatal if the price in the shopping cart or even at checkout did not adjust and the cached price was permanently displayed.
Such dynamic pages are therefore excluded by default in our caching to ensure proper operation. Simultaneous visits to websites with a shopping cart therefore usually hit the processor directly.
Communities and membership websites with forums and many people logged in are also difficult to cache. Here is a serious need for action in terms of hardware resources.
The Chips Shop as a meaningful Metaphor
Without caching, the only thing that helps is a higher number of CPU cores, which can then answer a high number of concurrent visitor requests. There is the chips shop as a metaphor here to explain the principle:
Each CPU core stands for one person behind the counter of the chips shop. The more people fry at the same time (cores are available), the more fries can be sold (requests can be answered).
By the way, this does not apply to the speed at which people work. For this, the CPU clock frequency would have to be increased. This is the case with our rates from PRO tariff, where the processors are up to 30 percent faster.
Approximately 600 visits simultaneously in the Shopping Cart as a maximum
With a two days' notice, we can manually upgrade tariffs to 36 cores and 64 GB RAM for 24 hours. This is especially popular with many startups that will have an appearance on TV shows like "Dragon's Den" and want to be on the safe side with the expected load peak.
For the large WooCommerce store there is our largest high performance tariff "Business XXL" with 12 vCores and 32GB RAM. Here can be up to 600,000 cached visits per minute or up to 600 visits per second in the shopping cart.
Load Balancing as the next step
In the event that even more visits per second need to be processed, load balancing is the next step. A load balancer is connected in front of the actual servers, which distributes the requests among the servers. Load balancing is therefore about distributing the load.
This allows scaling across hardware resources not only on the same server, but also across server boundaries. This method has been established for decades and is excellently suited to directing high traffic into sensible paths without downtime.
Load balancing can be implemented, for example, using a content delivery network (CDN). With a CDN, the servers are distributed worldwide and enable fast access to the website from almost any country. Each of these servers stores a cached version of your website and delivers it to the end device. A CDN is especially worthwhile when it comes to an international website, as not only the server load but also the latency plays a role here. Providers like Cloudflare offer the possibility of a CDN.
Conclusion - WordPress & High Performance already fit together very well
Who nowadays still classifies WordPress as a CMS for small websites, should urgently rethink. WordPress has long been established as a CMS for high-performance websites and is used in many ways. Especially caching helps WordPress here to true heights.
Even big online stores with a constant high traffic can be reliably handled with the appropriate hosting. In countries like the USA, this is not a secret anymore for a long time. Hopefully, I was able to show you in this article that we can also trust our favorite CMS in Germany.
How have you perceived WordPress so far? Have you built any high traffic websites? I look forward to your comments!