Our Obsessive Approach to Caching
Site updates appear in a flash with our cache process
When it comes to growing your site traffic and audience, we all tend to think of the same things: great content, beautiful design, fast load times, interactive experiences, etc. But one thing that often gets overlooked is the role of caching. A web cache is the temporary storage of web documents, like images, article pages, and multimedia assets, to reduce unnecessary repeat loading from a server. Any platform that doesn't focus on caching performance usually has poor uptime stats, which means that you're going to experience downtime issues.
RebelMouse obsesses over these things so that you don't have to, and that's why we're able to handle massive traffic spikes without any degradation in service. Just ask RebelMouse-powered PAPER Magazine, who recently had an article receive 20 million page views in one day (!) with no noticeable change in site performance.
The way we handle our cache process ensures that your editorial changes go live very quickly on your sites. It's just another step in our devotion to keeping our platform — and your sites — the very best in the world.
Here are three reasons why our cache approach works:
We have a five-minute cache time, and we're experimenting with increasing it to lifetime so that the cache never expires except when content has changed. Currently, we have an API integration with our CDN that clears the cache when a post is updated. We can also update our five-minute cache time to six-hours because we already have long TTLs for the sites we power. TTL, or Time to Live, is part of the authentication protocol that verifies a client's certificate with web servers. For caching assets, such as images or CSS, we use a one-year cache time.
Front-end Web Servers: We use Varnish servers installed on our front-end web servers that provide us with traffic-spike protection and improve our ability to serve responses to our CDN faster. For the most part, the configuration of our Varnish servers repeats what we have set up in Fastly.
Application: RebelMouse's application is tightly coupled with memcached, which is an open-source caching system designed for sites that have a lot of data. By having this paired with many of our own data structures that require heavy, CPU-bound processing, we can reduce both the cache population and prewarm time.
By keeping our caching rules so strong, we're able to maintain an incredibly stable and fast platform. It's one of the reasons we have so many amazing case studies. We're also able to avoid any penalties from social sites, like Facebook and Google, who will notice things like downtime and make your content less discoverable.