1

I have some difficulties wrapping my head around the following concept:

When HTTPS inspection (outbound) is enabled on a NextGen Firewall (NGFW) e.g. CheckPoint, a certificate has to be installed on the NGFW. These certificates (or at least the root/intermediate) need to pushed or installed on the clients to be able to "trust" the NGFW's certificate.

How can HTTPS inspection be successful if, for example, a browser sees that the Subject of the certificated (installed on the NGFW) does not match the Domainname/URL of the requested website?

In short: even though the certificate of the NGFW is regarded as trusted by the client on a CA-level, it still sees that it is not connected to the server it is intended to reach and therefore recognizes a MITM attack. Or am I missing something?

Kitetsu
  • 48
  • 6

1 Answers1

1

The certificate used in TLS interception is not a server certificate but a CA certificate. It is not used to authenticate the intercepted server itself. Instead a new certificate will be dynamically created for each visited server and this new server certificate will be signed by the intercepting proxies CA certificate.

In the client the dynamically created server certificate will undergo the usual checks, i.e. matching subject, expiration etc - all fits because it was specifically created as a replacement of the original sites certificate. The trust chain fits too since the proxies CA was explicitly imported as trusted into the client.

Steffen Ullrich
  • 184,332
  • 29
  • 363
  • 424