HTTP/2 Push allows a web server to send resources to a web browser before the browser gets to request them. It is, for the most part, a performance technique that can help some websites load twice or thrice as fast. ShimmerCat collects data, analyzes it and provides rules on how to push assets before the request is made. It can predict what the visitor needs based on conditions like geographical location, device, and browser. By doing so, assets can be pushed before the HTML file has been produced. With HTTP/2 Push you can use the time-to-first-byte (TTFB) - otherwise spent on waiting - to deliver website resources faster to your visitors. The web normally works by pulling assets: as soon as the browser knows that it will need a script, stylesheet or image, it pulls it. HTTP/2 Push allows the server to take the initiative by sending assets it knows the browser is going to need ahead of time. The challenge is making good use of HTTP/2 Push in the way ShimmerCat can do thanks to the AI engine that analyses data and create optimisation rules. In the video below you can get an easy-to-digest explanation of HTTP/2 Push compared to the traditional approach for delivering website resources ShimmerCat uses HTTP/2 Push in conjunction with cache digests and asset prediction to automatically reduce loading times in a way that no other web server does. How does HTTP/2 Push work with caching? The short answer is that it works well: if a resource is in the browser’s cache, there are ways to avoid pushing it. Our implementation of HTTP/2 Push is very good at avoiding reiterated push. Waterfall view Below you can see the effect of HTTP/2 Push in a waterfall view: For a more detailed technical deep-dive into HTTP/2 Push, check out our blog post A closer look to HTTP/2 Push.