Verifying SSL site certificate verification
-
Hi, I am using Pfsense with Squid, Squidguard and Snort packages. Squid with non-transparent proxy. I have set CA certificate (i.e. named "mycert"), when I open i.e. bank website (or other site with SSL) I see green padlock in my web browser, but it is only that green padlock, without any information about that who is verified certificate (i.e. Verysign, Symantec or other). There is only information, when I click that padlock, so cert was verified by Pfsense (mycert). How can I check that informations?
Sorry for my english and regards. -
With SSL MITM you can't. If you want to verify certificates (and who doesn't?), don't do that.
-
But SSL MITM is more security (in my opinion, if I'm not wrong?), how to otherwise filtering https sites with proxy?
-
Well, then you have a choice to make.
Look at https peek/splice
-
More security for who exactly? You can filter users from going to sites you don't want them to go to without having to do a MITM on the ssl. If your doing a MITM the user would not be able to see the end cert, only the cert you present them.
-
Thank you very much for a explnation, but can you tell me how can I filter sites with SSL and the same time have anti-virus protection enabled for them?
-
You would not be able to scan for antivirus inside a ssl tunnel. The client would do this on their end.. You wouldn't do this on the firewall.
But you can filter what sites they go to without having to do a mitm on the ssl.. Doing mitm opens up a whole can of worms.. You would then be able to sniff users passwords to like their banks, and such. View medical info they might be looking at.
SSL is suppose to provide end to end security.. From the end user to the site they are going to.. When you break that trust by doing a mitm you can open up a whole freaking can of issues that are best left alone..
-
So, if you can (please), tell me, what should I do to do? Use transparent squid proxy without MITM? And how to manage ssl sites (i.e. how to block youtube, facebook etc. for some users in my network)?
Best regards and thank you very much. -
peek/splice, not MITM.
-
Then what should I to do? Disable MITM, enable transparent proxy in squid? I found only in Squid (in SSL Man In the Middle Filtering section) that information: Please see SslBump Peek and Splice wiki documentation for additional details - but there http://wiki.squid-cache.org/Features/SslPeekAndSplice is not information how use it with Pfsense. Can you tell me some more about how do that, please?
-
Here is mine
-
-
Thank you, doktornotor. I have installed end-point antyvirus on every host too. In my network I must have any security solution (like IPS/IDS - I use Snort package), that why I used MITM and SSL filtering. Chrismallia show his configuration, but if I am not wrong he has MITM and SSL filtering too. Only different is with SSL/MITM mode - I have default, he Splice All. That's all?
-
with splice all you do not need to install certs on devices, it is mostly good for only filtering no AV, so it does not break the ssl connection or inspect content, it uses the SNI, anyone correct me if I am mistaken in any way
-
Thank you, doktornotor. I have installed end-point antyvirus on every host too.
Oh noes. That was NOT the point. The point was that MITM is evil.
-
I tried this settings (from Chrismallia) on virtual network, Pfsense and two hosts and what can I say… it works. Antivirus dont works with SSL, but I have it installed on every host. Thank you all for help and explanation very, very much. But one thing more that I can't understand (after Doktornotor post) - it is still MITM enabled - does it matter?
Best regards -
no with splice your not your doing a peek and then splicing it back.. The client is getting the actual cert from the server.. Your just peaking inside to see where they are going, etc. And if it should be filtered.
"splice: Become a TCP tunnel without decoding the connection. The client and the server exchange data as if there is no proxy in between."
-
Thank you very much again, so now I have everything ok. With Squidguard ssl filtering just works.
-
There are many attack vectors to infect machines when MITM is used. SSL provides two things, privacy and security. The security is provided by the signing. When you MITM, you break the signing, which means you break the security.
-
With MITM you're breaking the chain of trust normally provided the trusted third party signatures found on the certificates. Your users will no longer be able to verify that they are connecting to the real deal when they connect to a HTTPS site, instead they can at maximum be sure only that they are connecting to your MITM proxy.
