SnortSam NetScreen? Configuration

Comments, questions, additions, input and constructive criticism are always welcome.

Please read the entire file and perform a backup of your configuration before making any changes on your Netscreen. If you don't know how to perform a backup of your configuration, please obtain a users manual. Just a hint, a 'get config' command gets the entire configuration from your unit so if you can telnet into the Netscreen that is the command you need to run. Take the output on your screen and save it to a text file. We don't assume any reasonability if you blow up your Netscreen using these commands. We are just trying to help you get started and we can't take reasonability for anything that you did, haven't done, or are going to do. So, don't blame us, me or anybody on the SnortSam team.

SnortSam Configuration

Usage for the Netscreen option in the config is:

netscreen <firewall-ip> <login-id> <login-pw> <optional-group-name>

Netscreen Configuration

In order to get this working there are a few things that must be setup first on the Netscreen. The first thing is to setup a group on the Netscreen. The default name for the group is called 'SnortSam'. That is unless you change it in the conf file. The way SnortSam interacts with the Netscreen is that is adds and removes IP addresses to and from this group on the Netscreen.

Note: We picked groups instead of adding and deleting rules to the policy because of the way Netscreen does the rules and how they are added, and then moved when doing them from the CLI. If you are using routed mode you might need more than one rule to do blocking, depending upon if you have any inbound services. A group instead of adding rules would be a better fit. Keep in mind that the Netscreen, like other firewalls, reads the rules from top down. So you want to keep these rules that we are entering closest to the top, so that they can be most effective.

Routed and Transparent Mode Considerations

There are two modes that your Netscreen can operate in, routed mode and transparent mode. Routed mode is when you have separate layer 3 networks on each interface. Your trusted interface has two modes that it can use, but only if you are in routed mode. Those two modes are NAT and route. If you are using private addresses on your trusted interface then it should be set to NAT mode. Please note that if you are using NAT mode there are special considerations to make. Since there is only one group support for SnortSam right now, the best that we can do is block traffic going to your DMZ if you have one. Keep in mind that there is a default deny rule on your incoming policy, so blocking there using SnortSam really doesn't do anything. If you have any one-to-one network address translation going between your trusted and untrusted network, you will need to pick which is more important to block (for now). If you want to block on your untrusted interface like the examples below you will need to create a policy with an explicit deny for each inbound service. SnortSam will add the IP addresses to the group. The Netscreen 5's or 5xp make great screening walls for relatively low speed connections. I am talking about less than 3 meg links. They are cheap and you can still get Netscreen 5's.

If you are using transparent mode then you have dropped the Netscreen in between your internet router and your hosts or firewall. That would mean that you have a single IP address for management of the box. Unless you mess up the default configuration, nobody can get to the Netscreen from the outside in transparent mode directly. This, in my opinion, is a great way to do a screening wall. The best configuration is to put your IDS sensor between your screening wall and your firewall where SnortSam is doing the blocking, because you will be blocking everything before it hits your firewall. The configurations below assumes that you are using transparent mode but if you want to use routed mode there will be a few things that you will need to change.

Netscreen CLI commands (Basic)

The commands below assume that you are using transparent mode. If you are not using transparent mode, you have to keep in mind that the commands will need to modified to work. Once you have established a telnet session into the Netscreen, you should see a -> prompt. The first command is to create a group. If you can't telnet to the Netscreen, you must make sure that you can, or else SnortSam will not be able to block. Go to the web GUI, and under the interface tab go to 'trusted'. Make sure that telnet is checked. You might need to reboot the Netscreen in order for that to take effect.

Group Command:

Set group address untrust "SnortSam" comment "Put your comment here"

Now, the second command depends upon how many incoming rules you have. If you do a 'get config' at the Netscreen CLI prompt, and look at the set policy commands, you will see the rules that are currently being used. Note that the rules are not in order by their id, but in the order of how they are enforced. Remember, they are read from the top down and the id number is only used to reposition the rules. So if your policy rules read like the ones below, you will need to insert a rule before id 1. We don't care about the outgoing policy for now, just the incoming policy.

Partial 'get config': set policy id 0 outgoing "Inside Any" "Outside Any" "ANY" Permit log set policy id 1 incoming "Outside Any" "Inside Any" "ANY" Permit

Policy Rule Command: set policy before 1 incoming "SnortSam" "Inside Any" "ANY" Deny

By default, the group name 'SnortSam' is used. But if you want to use something else you can do so via the SnortSam configuration file. Note we use the 'before 1 incoming' in the rule above. But if your first incoming rule is id 5, then the command would be 'before 5 incoming'. You can add the following after the deny if you want, but keep in mind all these take up more memory and processor resources.

count enable counting log enable logging schedule set scheduler traffic traffic parameters

After you have executed the command, your policy rules should look like this:

set policy id 0 outgoing "Inside Any" "Outside Any" "ANY" Permit log set policy id 2 incoming "SnortSam" "Inside Any" "ANY" Deny set policy id 1 incoming "Outside Any" "Inside Any" "ANY" Permit

That is it for the Netscreen configuration.

Known Issues:

Some of the Netscreen's, depending upon the code version that you are using, may only let one person from the same IP address telnet into the box. So, if you are testing this out and want to see the modifications in real time, either use the web interface or a different box to view the modifications.

This is more of an FYI than an issue but with the new code 4.x there might be some issues with this configuration. I am just saying this because I don't know, I haven't tested it. I should have a chance to test it in a week or two. The structure in 4.x is a little different so beware.

Regards, Christopher Lyon cslyon@netsvcs.com

-- MattJonkman - 09 Mar 2007

Topic revision: r1 - 2007-03-09 - MattJonkman
 
This site is powered by the TWiki collaboration platform Powered by Perl This site is powered by the TWiki collaboration platformCopyright © Emerging Threats