Most people lock their doors at night, however if you walk into someone's home you likely won't find every piece of furniture bolted to the floor as well. We trust that if someone is inside our home they are supposed to be there. Unfortunately many developers treat local networks just the same, assuming all internal HTTP traffic is trusted, however this is not always the case. They incorrectly assume that their services will be protected by the same-origin policy in browsers, rather than implementing proper authentication mechanisms. By abusing this implicit trust we can gain access to confidential data and internal services which are not intended to be publicly accessible.
I will demonstrate that this is a poor security control and can be trivially bypassed via an older technique, DNS rebinding. The talk will cover how DNS rebinding works, the mitigations imposed by modern browsers and networks, and how each mitigation can be bypassed. I will discuss the notorious unreliability of DNS rebinding attacks that causes many developers to ignore the issue and how to overcome this unreliability.
Finally, I will examine a variety of popular services and tools to understand how they are affected by DNS rebinding. I will be releasing a tool that allows researchers to automate DNS rebinding attacks, the associated mitigation bypasses and generate drop-dead simple proof-of-concept exploits. I will demonstrate this tool by developing exploits for each vulnerable service, ending the talk by exploiting a vulnerable service to obtain remote-code execution, live.