You’re working on a new project, which uses a framework requiring strong random number generation. Typically, this could be Spring Security’s BCryptPasswordEncoder. You run your application on a Linux machine: your local Linux Mint for development or some Ubuntu Server instance, for example. And then… nothing happens. The application seems to be hanging! What gives?
Previously we have managed to finally get rid of that Session object and transform our RESTful API security to leverage a stateless authentication solution. We replaced our beloved JSESSIONID with a simple string of text, a token, that allowed us to identify a user. But the solution we used is not secure at all. We still need to find a token with the right characteristics to be safe enough in this world of leaks and breaches. Enter JWTs…
Last time we reviewed how to quickly set up stateful authentication on our Spring-based project. That’s very nice ‘n’ all, and in many cases you won’t need anything more. However shouldn’t we try to get rid of that session-based dependency and attempt to move to a REST-friendly stateless authentication solution? Let’s begin…
After a quick introduction we are now ready to begin our journey towards stateless authentication for RESTful APIs… by setting up a stateful example. Yes I know, but we have to start somewhere, right? In this part we’ll set up our project and code a couple of simple endpoints. One of those will be secured using Spring Security’s session-based authentication.
By now “stateful” or session-based authentication is pretty much well-accepted. Frameworks such as Spring Security or Apache Shiro make it really easy to implement a decent solution in just a few easy steps. I previously discussed how to secure a Spring-based REST API using Spring Security for authentication, CSRF protection and CORS. But in some cases, session-based security might not be good enough…