I’d like to start posting more on my website. One of the ideas I had for doing that is to write about some of the more obscure things in application security. My career has been almost entirely on the job training. This comes with the challenge of trying to understand explanations that sometimes don’t state obvious ideas and concepts. I’d like to cover some of those obscure topics and really dig into them. In the process I’d like to try and answer them in a way that my younger self can understand. I’m hoping this will provide others with a simpler explanation while helping me to better understand the concept.
To kick it off, I’d like to cover pragma and cache-control. The reason why is because I see this regularly in reports from ZAP, “Incomplete or No Cache-control and Pragma HTTP Header Set.” It’s a low or information finding and it usually shows up in a lot of places. I did some research to better understand the finding several weeks ago.
What is pragma and cache-control?
Both are header settings for the browser. Before we get to far, I guess I should explain what are HTTP header settings. These are settings that the application will communicate to the browser to determine how they interact together. In the case of pragma and Cache-control the application is telling the browser not to cache any content or don’t keep any content pulled down in the browser after leaving the site. The reason is so that sensitive information doesn’t get stored in the browser for a malicious actor to take advantage of if the person’s computer is compromised. For a more in depth explanation on HTTP headers check out this article, HTTP Headers for Dummies.
Pragma is an older header setting for HTTP/1.0. Cache-Control is the newer header setting for HTTP/1.1. What that means is that pragma is used to control cache for older browsers. If older browsers are being supported by the website or application then setting that is important to ensure sensitive information isn’t being stored. If older browsers are not support then just cache-control is used. In some cases you may have both. The Mozilla Developer Network has a pretty good explanation. There’s also a good explanation and discussion on Stack Overflow if that’s your thing.
How and where to set pragma and cache-control?
Where the header settings will depend on the application and what is being sent as content to the browser. The only setting for pragma is no-cache. Cache-control has more directives to set for flexibility. Max-age and revalidation are two directives that can be set. The browser will have to stick to those settings. The recommendation I’ve seen is to set pragma with no-cache and cache-control with no-cache, no-store, must-revalidate, and private.
These settings are set in either code or on the server. How implementation is done depends on the language and technology used by the application. Google is a good place to look for how to set these settings based on those factors.
Simple as that.
Hopefully, this clears up some confusion for people. Writing this down has helped me understand it a little better. I hope to do more of these types of articles in the future. I would love feedback on this post. Also feel free to drop a comment if I missed something in the explanation.