How to Know When You Are Being Probed Intrusion Detection for FW-1 Lance Spitzner Last Modified: 23 May, 2000 Firewalls do a good job of keeping the bad guys out. But wouldn't it be nice to know when the bad guys are knocking on your door? This article covers just that, how to determine when the bad guys are probing your network. We discuss how you can use the FW-1 User Defined Alert to track when you are being probed, and by whom. If you would like to see actual intrusion detection results, click here. NOTE: This script works only on Unix based systems. Download the latest version here. Ports: The IDS script has been ported to run on NT systems. I have not personally tested any of the scripts, you are on your own. James Oryszczyn has developed detect.pl, written in PERL Oscar Wahlberg has developed fw1_alert.pl, written in PERL Roberson Andrew has written alert.zip, written mainly in Windows Script Host 5.1 New Features with ver 1.4.3: Removed TrackDown() function, due to changes in Whois database. Tested and runs on most Unix flavors, including Linux, Solaris, and Nokia Other Features Include: An installation script that builds, installs, and tests everything for you. Includes an optional feature that automatically determines the the admin of the scanning source and notifies them via email. Also new with FW-1 ver 4.0, you have the option to automatically block systems that are scanning you. You should also find this version much easier to read and understand. You can add your own alerting modules, such as page alerts or snmp traps. Download the latest version here. The Problem How do you know if someone is probing your network? Logs are a great place to start, especially your FW-1 logs. Unfortunately, those logs can be difficult and time consuming to sift through (especially if you are logging 400+ MB every day). Wouldn't it be great if there was an automated method of alerting you? Well, this article does just that. There are a variety of different probes and attacks black-hats will attempt. The type we will be focusing on is port scans. Port scans are where an individual attempts to connect to a variety of ports to identify what services a system is running. The scans can be used on a specific target, or used to scan entire IP ranges, often chosen at random. This is one of the most popular information gathering methods used by black-hats today as it identifies what ports and services are open. If you are interested in learning more about these techniques, check out Know Your Enemy. The Goal The goal is to receive an email alert whenever someone is attempting to scan or probe your network. We want to automate the entire process, saving you time and energy. We also want to archive this information in a database for future use. Our definition of an email alert would look as follows: Date: Wed, 31 Dec 1997 15:40:01 -0600 (CST) From: ids@example.net To: fwadmin@example.net Subject: #### Firewall ALERT #### You have received this message because someone is potentially scanning your systems. The information below is the packet that was denied and logged by the Firewall. This is email alert number 3, with a limit of 5 from evil.example.org. ----- CRITICAL INFORMATION ----- Date: 31Dec1997 Time: 15:39:59 Source: evil.example.org Destination: ns1 Service: domain-tcp ----- ACTUAL LOG ENTRY ----- 31Dec1997 15:39:59 drop fw1 >elx0 mail proto tcp src evil.example.org dst ns1 service domain-tcp s_port 37401 len 44 rule 6 As you see, this email is alerting the fwadmin that someone has attempted to conduct an un-authorized zone transfer from the nameserver. However, the firewall denied the attempt, logged it, AND alerted the fwadmin. In the email alert, we have parsed the critical information for easy reading, and also included the actual log entry for more data. You can also creat your own alerts, such as text pages or snmp traps. The Solution To detect and monitor these scans, we will build a script that first logs whenever someone connects to a predetermined port, and then emails us an alert. To do this, we will be using the User Defined Alert. The User Defined Alert will be our script, alert.sh. Whenever the User Defined Alert is activated, it will launch our script, which parses the data and does all the work for us. We set up our Firewall rule base as follows.. First, we identify the most commonly scanned ports we want to monitor. Then we select several systems to listen on these ports. When an intruder scans our network, he will most likely hit our systems listening on these ports. When these ports are scanned, the Firewall logs the attempts then executes the User Defined Alert.. The end result is you receive an email for each port scanned. If you have 3 systems, each listening on 4 ports, then you may get up to 12 emails from a single network port scan. However, this is normally not the case. If black-hats are scanning an entire network, they are normally looking for a single vulnerability, such as imap (port 143). In this case, we would have received only three emails, one from each system. When they scan a single target, often they scan a range of ports, such as 1-1024. In that case, we would have received only 4 emails, one for each port on the system. Based on what emails you get, you can quickly determine what the intruder is interested in. Also, the script alert.sh is configurable to limit the total amount of emails as a result of a scan (the default limit is 5). All scans are also logged to a database. For those of you with more then one firewall, the script should be installed on the Management Server. The Management Server receives all firewall logs and executes the User Defined Alert (our script) local on itself. So if you have 5 Firewall modules reporting to a single Managment Module, you would install the script on the single Management Module. All 5 Firewall modules can use the script(as defined in their rulebases), however the script is executed on the Management module (where all the logs are). This is extremely advantageous to you, as you can detect scan patterns happening across multiple firewalls. Implementation To implement this methodology, we first identify two to three systems for each firewall to use for monitoring. I often select DNS servers, as these are primary targets, many scanning tools start by scanning these systems. Once you have identified the systems you will monitor, create a workstation object in your Firewall 1 GUI for each of the systems. The systems do not necessarily have to actually exist. Create an object for a non-existing system, give it a valid IP that your are not using, and use this as one of your scannable objects. Once you have created your workstation objects, then select the most commonly scanned ports. Ensure that your systems are not using these ports, or every time someone connects to it, you will be alerted. To identify commonly scanned ports, CERT alerts are a great place to start, you can find these alerts at http://www.cert.org. Some ports that you can use are: dns zone transfer 53/TCP portmapper 111/TCP http 80/TCP SMB 139/TCP imap 143/TCP BackOrrifce 31337/UDP --> Yes, people are STILL scanning for old BO. I like these ports since black-hats commonly look for them, but most of your systems will not be using them. Make sure these ports are not already blocked by a screening router. We will then set several system to listen on these ports, alerting us when there is a connection. Below is an example of a rule that monitors and detects the scans. Notice that quite a few ports are being monitored for these systems. [Image] As you can see from the image above, this rule base is monitoring quite a few ports. Also, this rule is our User Defined Alert rule. Whenever this rule receives a connection on one of these ports, the User Defined Alert will be activated. Our User Defined Alert (alert.sh) will now parse the data for these connections, notify us via email, and archive the information. Be sure to configure your User Defined Alert to point to the alert.sh script. You do this by going into Configuration Properties, Alerts, and then giving the full path to the alert.sh script (such as /home/fwadmin/alerts/alert.sh) Below is another example of how to use the alert.sh script. Your DMZ should never initiate traffic to your internal network. By design, you do not trust your DMZ since it is accessible by the Internet. So, I created an additional rule to log and alert me whenever the DMZ attempts to initiate a connection to the internal network. If the DMZ does attempt to reach your internal network, it may mean one of the DMZ systems has been compromised and the bad guys are trying to get in. Notice how I put this rule just before the last "catch all and drop" rule. [Image] That's it, you are done. All you have to do is download the latest version and run the installation script. If you want to learn how the script works or modify it, read below. --- This is informational only, you DON'T have to read this --- The whole alerting process works as follows. First, we define the User Defined Alert (shown in the image above) as our alert.sh script.. The User Defined Alert produces a log entry for every connection that it monitors. The log entry looks as follows: 31Dec1997 15:39:59 drop fw1 >elx0 mail proto tcp src evil.example.org dst firewall service domain-tcp s_port 37401 len 44 rule 6 The Firewall then pipes this log entry into our alert.sh script. The script then parses the data and executes various functions we define. The script alert.sh uses the following files to store and track the logs. alert.uniq - A single, permanent log is added to this file whenever a system scans you. Regardless of how many ports are scanned, there will only be one entry per unique source. This way you can easily track who scanned you when. alert.log - This file temporarily stores every port scanned for that day. This is how the alert.sh script tracks how many times/ports have been scanned from a single source. Every time the User Defined Alert executes alert.sh, the first thing the script does is count how many times the source ($src) has scanned us for that day by looking in alert.log. When the script alert.sh is done, it adds an entry to the alert.log file. So, the first time a source scans a port, alert.sh will not find anything logged in the alert.log. The second time the sources scans a port, it will find one, and so on. You set a limit on how many ports are scanned before the alert.sh stops reacting (this protects against DOS attacks). At the end of the day, a cronjob (rotate.sh) will clear this file out and add all the entries to archive.log. This way if the same source scans you again at a later date, alert.sh will log the attempt. alert.archive - This file permanently archives every port scanned and logged. This is an archivial database that can be used for research. This file stores the actual log entries, so it tends to be difficult to read. I find the file alert.uniq to be much easier to read. rotate.sh - This is a cronjob you need to run every day. It clears the file alert.log, saving all the logged entries to archive.log. This resets the counter for the alert.sh script. If the same source scans you two days in a row, this allows you to log it. To get a better understanding of how our alert.sh script works, read below. The script is broken down into 6 parts. 1. First, we need to customize our script, which is done in the beginning of the script. This should be self explanatory. ################################################################# # BEGIN CUSTOMIZING SCRIPT HERE # ################################################################# # INSTALL DIRECTORY # Define the directory that this script is in. # Do NOT put a slash at the end. # EXAMPLE: dir=/home/fwadmin/alert dir= # FW ADMIN # Define the name of who gets the email alerts # EXAMPLE: user=fwadmin@example.com user= # SCAN LIMIT # Define maixmum number of scans/email alerts limit=5 # EMAIL REMOTE SYSTEM # Define as "true" if you want to automatically email # the remote admin when you reach your scan limit. email=false # SAM # Define as "true" if you want to autotmatically block # the source if you reach your scan limit. sam=false # SAM TIMEOUT # How long do you want the source blocked # Default is 3600 seconds (1 hour). timeout=3600 ################################################################# # FINISH CUSTOMIZING SCRIPT HERE # ################################################################# 2. We then define our system variables. Also, we check to see if we have met our scan limit. If we have, bail now to save CPU cycles. This is where the speed optimization happens. ### Script variables message=/tmp/.message_$$ send=/tmp/.send_$$ ### Good code is secure code umask=177 PATH=/usr/bin:/sbin:/usr/sbin:/usr/local/bin export PATH if test -a $message then rm $message fi if test -a $send then rm $send fi ### Set trap in case of abrupt exit trap "rm $send $message ; exit 5" 1 2 15 ### Grab User Defined Alert log, pipe to $message. cat - | tail -1 > $message ### Determine number of scans. ip=`awk '{print $10}' $message` number=`grep -c $ip $dir/alert.log` scan=`expr $number + 1` ### Check number of scans. If we have reached our limit, lets bail ### now and save CPU cycles. if [ $scan -gt $limit ];then cat $message >> $dir/alert.log rm $message exit 10 fi ### Parse log file date=`awk '{print $1}' $message` time=`awk '{print $2}' $message` dst=`awk '{print $12}' $message` ### Determine service (check some variables first) #Determine if "Valid Address" is in log files for NAT nat_check=`grep -c "(Valid Address)" $message` #Determine if protocol is icmp icmp_check=`grep -c " icmp " $message` if [ "$nat_check" -eq 0 ];then if [ "$icmp_check" -eq 0 ];then service=`awk '{print $14}' $message` else service=`awk '{print $15,$16,$17,$18}' $message` fi else if [ "$icmp_check" -eq 0 ];then service=`awk '{print $16}' $message` else service=`awk '{print $17,$18,$19,$20}' $message` fi fi 3. We then define all functions / modules used in the script. The script comes with three predifined modules, you can add more of your own. * Resolve () - If the source is an IP, resolve it to a fully qualified host name. * Block () - Blocks any remote system that reaches its scan limit. Works on FW-1 ver 4.0 only. 4. We then build our alert emails. This is the template used for every email alert that is generated. cat < $send You have received this message because someone is potentially scanning your systems. The information below is the packet that was denied and logged by the Firewall. This is email alert number $scan, with a limit of $limit from $src. ----- CRITICAL INFORMATION ----- Date: $date Time: $time Source: $src Destination: $dst Service: $service ----- ACTUAL FW-1 LOG ENTRY ----- `cat $message` EOF 5. Last, we reach the phases part. This is where we define what the script does (in addition to the alert emails) based on how many time the source has probed us. This is broken down into three parts, or phases. Feel free to change/modify these phases as you see fit. I've tried to make this as modular as possible for just that reason. Regardless, once we are done, we send off the email alert. * PHASE 1: This is when we receive the very first packet. The entry is logged in the alert.uniq file. * PHASE 2: This is when we receive the 2nd to $limit-1 packets. Currently nothing happens during phase 2 * PHASE 3: This is when we receive our $limit packets (default is 5). Here is where most of the action happens. This is where the functions Blocked(), TrackDown(), and anything else you want to add. ### Send email alert and save log to alert.log /usr/bin/mailx -s "#### SCAN ALERT ####" $user < $send cat $message >> $dir/alert.log 6. Don't forget, at the end of the day, run the cronjob rotate.sh. The purpose of this cronjob is to take all the logged entries from alert.log and move them to the permanent archive file alert.archive. This also clears out the alert.log file, allowing it to start counting new sessions. Thats it! I hope this does not sound complicated. Once you get it up and running, you will see how simple it is. I hope this helps you in your intrusion detection. While this is not the ultimate end-all solution, it is a great and easy start to automating your intrusion detection. If you have any recommendations or corrections, I would love to hear from you. Please email me at lance@spitzner.net. Author's bio Lance Spitzner enjoys learning by blowing up his Unix systems athome. Before this, he was an Officer in the Rapid Deployment Force, where he blew up things of a different nature. You can reach him at lance@spitzner.net . Whitepapers / Publications