Link to home
Start Free TrialLog in
Avatar of jameswright1337
jameswright1337

asked on

How to catch web proxy browsing using Snort/Squil

I am one of the network analyst with my company and I have been seeing an increase in web proxy use and I'm hoping to get some kind of help. Right now the only way I can find someone browsing the Internet via a proxy is by taking the IP address of a known proxy site and running a sancp query for it, to look for any traffic for the past few days. Sounds tedious I know, but we can't turn on the HTTP preprocessor in Snort because it causes way too many alerts, almost all false-positives. Snort hasn't updated that preprocessor in quite some time, so how it finds http_proxy is by looking for a URL within another URL, which is how ad banners and pop-ups work nowadays. We could write a rule to look for any instance of the word 'proxy' or 'anonymous', but again, we'd get way too many false-positives and not all web proxy sites have those words within the page. While researching how exactly proxies work, I found an article that stated that anonymous proxies do not send the HTTP_X_FORWARDING_FOR variable to the host. Not quite sure what I could do with that info (even our Snort engineer doesn't understand how to use that within a rule). I tried looking at multiple transcripts from a proxy hit and from normal http traffic, but I cannot for the life of me find anything that distinguishes one from the other. We also use Smart Filter which uses the database from securecomputing.com, but that thing goes down all the time, making it virtually useless.

Any help with this would be great, as I am hoping to catch the majority of proxy users rather than the measly 10% I get on good days. Thank you for any inputs you can provide.
Avatar of tellkeeper
tellkeeper

Complex. :) . Write a rule rejecting access to all requests lacking HTTP_x_forwarding_for.
Avatar of jameswright1337

ASKER

Due to upper-management problems that occurred before I got hired on, we are not allowed to block via Snort using the reject/deny rules. Maybe I can convince them though with all of the hits we are getting with proxy use. Any idea as to how the rule would be written? I've only had to write rules that looked for 'content:' in the package, with a few having 'depth:' and 'distance:'.
ASKER CERTIFIED SOLUTION
Avatar of mail2divyesh
mail2divyesh

Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
Sorry mail2divyesh, that's not exactly what I'm looking for.


Are there no Snort experts here at all? Guess I can't be too surprised, I couldn't get this answered on the Snort.org forums.
James,

Have you got a link to that article on HTTP_X_FORWARDING_FOR variable?

I could have a tinker with some rule sets to see if I can find a way to fit it in.

Otherwise, off the top of my head, dumping a list of those known proxies in to a snort log rule set and doing anomaly detection against user's web site choices (slow, boring and painful) 100+ hits against site x is the only way I can thing of with snort.



What90,

Can't for the life of me find the article, but it did say that if a connection is using a proxy, the fowarding_for will have a variable. It looks to be an old theory because I looked at probably 100+ transcripts through Ethereal and couldn't find a definitive commonality in proxy transcripts or any major differences between proxy and normal traffic. I think for right now I'm going to have to just keep vigil in scouring the Internet for new proxy sites and get them either blocked or categorized correctly in SmartFilter. Thanks for all the input.
SOLUTION
Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
That's what I've been doing as well. Any time I find a new proxy site, I make sure to have it categorized correctly at Secure Computing. If it is a SSL site, that's when I submit the block request to the firewall team. So far I've been able to limit the use of web proxies almost completely. I was just hoping that there might be a way to reject the packets altogether. Thanks to everyone who commented.