Intrusion detection, integrity checks: IDS, NIDS, HIDS, Antivirus, software.
Note: vulnerability checking: CIS, SATAN, COPS, Tiger
FAQ: Network Intrusion Detection Systems: http://www.robertgraham.com/pubs/net...detection.html
Sniffin' the Ether v2.0: http://www.unixgeeks.org/security/ne...r/sniffer.html
Lotek sniffing docs: http://www.l0t3k.org/security/documents/sniffing/
Defeating Sniffers and Intrusion Detection Systems, Phrack, http://www.phrack.org/show.php?p=54&a=10
The IDS acronym game:
IDS: Intrusion Detection System refers to an application able to examine traffic for attributes and properties that mark "benign", suspicious, restricted, forbidden or outright hostile activities.
NIDS: Network IDS refers to Intrusion Detection, like running "sensors" on various sentry or sniffer hosts while logging and/or logprocessing and alerting is done on a central host (many-to-one topology).
NIDS examples are:
OSSIM (Snort+Acid+mrtg+NTOP+OpenNMS+nmap+nessus+rrdtool): http://sourceforge.net/projects/os-sim/
Panoptis (DoS, DDoS only):
Some commercial/non OSS examples: Demarc PureSecure, Cisco Secure IDS (NetRanger), ISS Real Secure, Axent Net Prowler, Recourse ManHunt, NFR Network Flight Recorder, NAI CyberCop Network, Enterasys Dragon and Okena Stormfront/Stormwatch.
Snort also is available commercially these days.
HIDS: Host-based IDS. The HIDS acronym itself is subject to flamewars.
IDS examples are Snort, Shoki, Prelude, Defenseworx, Pakemon, Firestorm and Panoptis (DoS, DDoS only).
IPS: Intrusion Protection System. Passive or active (learning, like the heuristics stuff?) enforcement of rules at the application, system or access level. I suppose we're looking at stuff like Grsecurity, Solar Designer's Open Wall, LIDS, LOMAC, RSBAC, Linux trustees, Linux Extended Attributes, LIDS or Systrace here.
Commercial/non OSS examples: Entercept, ISS RealSecure, Axent Intruder Alert Manager, Enterasys' Dragon, Tripwire, Okena and CA's eTrust.
Intrusion Detection Systems: An Introduction: http://www.linuxsecurity.com/feature...story-143.html
Intrusion Detection FAQ (SANS, handling ID in general): http://www.sans.org/resources/idfaq/index.php
Basic File Integrity Checking (with Aide): http://online.securityfocus.com/infocus/1408
(IDS, NIDS, File Integrity Checkers)
Using Snort as an IDS and Network Monitor in Linux (SANS, PDF file): http://www.giac.org/practical/gsec/James_Kipp_GSEC.pdf
Snort: IDS Installation with Mandrake 8.2, Snort, Webmin, Roxen Webserver, ACID, MySQL: http://www.linux-tip.net/workshop/id.../ids-snort.htm
ArachNIDS (Snort/Dragon/Defenseworx/Pakemon/Shoki rule, research and info library): http://whitehats.com/ids/
Intrusion Detection and Network Auditing on the Internet: http://www.infosyssec.net/infosyssec/intdet1.htm
Snort Stealth Sniffer: Paranoid Penguin: Stealthful Sniffing, Intrusion Detection and Logging: http://www.linuxjournal.com/article.php?sid=6222
Dropping Packets with Snort:
Why not to use Snort's "flexresp": http://www.mcabee.org/lists/snort-us.../msg00379.html
Guardian: see the Snort tarball, in the contrib dir.
Snort GUI's, management, log reporting and analysis:
Snort Unified Logging: Barnyard: (Sourceforge)
Snort Unified Logging: Logtopcap
Snort Unified Logging: Mudpit
Analysis Console for Intrusion Databases (ACID): http://acidlab.sourceforge.net/
HOWTO Build Snort with ACID: http://www.sfhn.net/whites/snortacid.htm
ACID HOWTO: http://www.andrew.cmu.edu/~rdanyliw/...snortacid.html
ACID FAQ: http://www.andrew.cmu.edu/~rdanyliw/snort/acid_faq.html
SPADE, Snortsnarf: http://www.silicondefense.com
Enabling Automated Detection of Security Events that affect Multiple Administrative Domains: http://www.incident.org/thesis/book1.html
Demarc (commercial): http://www.demarc.com
Oinkmaster (rulemanagement): http://www.snort.org/dl/contrib/sign...nt/oinkmaster/
Snort alert mailer (C or .pe?r?l?): http://rouxdoo.freeshell.org/dmn/snort/
Pig Sentry: http://web.proetus.com/tools/pigsentry/
IDS Policy Manager Version (W32): http://www.activeworx.com/
Snort_stat: snort_stat.pl /var/log/snort/alert | /usr/lib/sendmail <firstname.lastname@example.org>
Swatch: ./swatch -c /root/.swatchrc --input-record-separator="\n\n" --read-pipe="tail -f /var/log/snort/alert" --daemon
Swatch + Hogtail.
Snort vs Abacus Portsentry:
Snort and PortSentry compared: http://www.linux.ie/articles/portsen...rtcompared.php
Comparison of IDSs ( NFR NID, Snort, INBOUNDS, SHADOW, Dragon, Tripwire): http://zen.ece.ohiou.edu/~nagendra/compids.html
Snort help, mailinglist (archives), honeypots:
Snort: Database support FAQ: http://www.incident.org/snortdb/
Snort mailinglists, Aims: http://marc.theaimsgroup.com/
Snort IDS forum at Whitehats.com: http://whitehats.com/cgi/forum/messa...?bbs=forum&f=4
Baby steps with a honeypot: http://www.lucidic.net/whitepapers/mcooper-4-2002.html
Honeypot & Intrusion Detection Resources: http://www.honeypots.net/
The TCP Flags Playground (Mailinglist, Neohapsis): http://archives.neohapsis.com/archiv...0-03/0386.html
Snort + 802.11 aka Wireless: http://www.loud-fat-bloke.co.uk/w80211.html
Sniffing (network wiretap, sniffer) FAQ: http://www.robertgraham.com/pubs/sniffing-faq.html
Apps, network monitoring (index): http://www.mirrors.wiretapped.net/se...ng-README.txt.
An Analysis of a Compromised Honeypot (Snort+Ethereal): http://www.securityfocus.com/infocus/1676
To add: Firestorm NIDS, Barnyard, Mudpit, Snort GUI's, add-ons etc etc.
Snort on two interfaces, solution one: "-i bond0".
Valid-for: running one Snort instance, multiple promiscuous mode interfaces except the mgmnt one.
Do once: "echo alias bond0 bonding >>/etc/modules.conf"
At boot: "ifconfig bond0 up; ifenslave bond0 eth0; ifenslave bond0 eth1"
At boot: start Snort with interface arg "-i bond0"
Snort on two interfaces, solution two: "-i any"
Valid-for: running one Snort instance, all interfaces.
Caveat: you loose promiscuous mode.
At boot: start Snort with interface arg "-i any" and a BPF filter to stop it from logging the loopback device.
File Integrity Detection Systems
Checking a filesystem's contents against one or more checksums to determine if a file (remember anything essentially is a file on a Linux FS) has been changed.
(for remote mgmnt see also ICU http://www.algonet.se/~nitzer/ICU/
or RFC http://sourceforge.net/projects/rfc/
which handles Aide, Integrit and Afick)
(for remote mgmnt see docs)
Tripwire (for remote mgmnt see FICC: http://freshmeat.net/projects/ficc/
Chkrootkit (not only Linux): http://www.chkrootkit.org
Rootkit Hunter (not only Linux): http://rkhunter.sourceforge.net
Commercial/non OSS examples: Versioner, GFI LANguard System Integrity Monitor, Ionx's Data Sentinel, Tripwire for Servers and Pedestal Software Intact.
File Integrity (SecurityFocus, tools list): http://www.securityfocus.com/tools/category/7
Viruses on Linux/GNU, Antivirus software
Sendmail, Tcpdump, OpenSSH, TCP Wrappers, Aide and some other projects have suffered from people succeeding to inject malicious code, and of those only Sendmail and OpenSSH where at main servers, the rest where mirrors AFAIK. Even though all the apps mentioned are safe to use, and the differences where noted soon, the real problem is you I. have to have the knowledge to read code, and II. the discipline to read the code each time and question any diffs or III. have minimal "protection" in place to cope with like rogue compiled apps "phoning home". Which in essence means to end users any SW provided w/o means to verify integrity of the code and the package should be treated with care, instead of accepting it w/o questioning.
As for the "virus" thingie I wish we, as a Linux community, try to "convert" people away from the typical troubles of Pitiful Operating Systems (abbrev.: POS, aka the MICROS~1 Game Platform) and direct them towards what's important to know wrt Linux: user/filesystem permissions, b0rken/suid/sgid software, worms, trojans and rootkits.
Basic measures should be:
- Using (demanding) source verification tru GPG or minimally md5sums,
- Watch system integrity (Aide, Samhain, Tripwire or any package mgr that can do verification: save those databases off-site, also see Tiger, Chkrootkit),
- Harden your systems by not installing SW you don't need *now*, denying access where not needed and using tools like Bastille-linux, tips from Astaro,
- Patch kernel to protect looking at/writing to crucial /proc and /dev entries and/or use ACL's (see Silvio Cesare's site, Grsecurity, LIDS),
- Watch general/distro security bulletins and don't delay taking action (Slapper, Li0n etc),
- Keep an eye on outgoing traffic (egress logging and filtering),
- Don't compile apps as root but as a non-privileged user,
- Inspect the code if you can,
- Don't use Linux warez,
But most of all: use common sense.
*If you're still not satisfied you've covered it all you could arm yourself with knowledge on forensics stuff like UML, chrooting, disassembly and honeypots.
If you want to find Antivirus software, Google the net for Central Command, Sophos, Mcafee, Kaspersky, H+BEDV, Trend Micro, Frisk, RAV, Clam, Amavis, Spam Assassin, Renattach, Ripmime, Milter or Inflex.
- AV SW is as good as it's signatures/heuristics. Some vendors don't update their Linux sig db's very well, or field SW with lacking capabilities. I've tested some (admittedly a long time ago) on my virus/trojan/LRK/malware libs. Bad (IMHO): Frisk's F-Prot (sigs), Clam (sigs), H+BEDV (libc version). Good (IMNSHO): Mcafee's uvscan (best) and RAV (2nd). Please do test yourself.
- AFAIK only KAV (Kaspersky) has a realtime scanner daemon. I'm in limbo about it's compatibility with recent kernels tho.
Links to check out:
LAVP/Mini-FAQ Linux/Unix AV SW
(list of AV vendors),