Slashdot Mirror
Is Tripwire Still Relevent?
Deagol asks "I work for a good-sized University. I've heard that Tripwire and our software licensing department is negotiating for a site-license. I was asked to comment on whether our department would like to buy in. I personally lost interest in Tripwire when they went commercial (I guess seeing a well-respected research tool go proprietary soured the milk for me), and though I've toyed briefly with the 'open source' version, I mainly have experience with the Academic Source Release. Seeing how their demo is only a 'simulation' (how lame is that?), I can't get a feel for what the commercial version can really do for me. Does anyone know the value (if any) of commercial Tripwire over the free one; Are there open source packages that have made Tripwire obsolete?"
Check out ViperDB, written in perl, it does it's checks every 5 minutes, in a highly optimized way at that. I actually know the guy who wrote it, and when setting up software on his machine set off his pager something fierce.
We use samhain. It's very nice because it can log to a remote host and store the filesystem database on a remote host as well. It also runs as a deamon and scans at a set interval. You can even make it change its name and hide its code in image files so as to trick hax0rs into thinking that its not installed.
The only thing I don't like about it is that I have it scheduled to check the machines every 10 mins, so if one of the junior admins changes something and forgets to reset the database I get an email every 10 mins until I reset it.
The homepage for samhain is http://la-samhna.de/samhain/
-----BEGIN GEEK CODE BLOCK----- Version: 3.12 GIT d? s: a-- C++++ UL++++ P++ L+++ E- W++ N o-- K- w--- O- M+ V PS+ P
is Osiris, which has an Apache-style syntax and a weird pseudo-free license. I haven't worked with enough filesystem integrity management systems (aka intrusion detection systems) to differentiate its use from Tripwire. My two cents.
I guess this will not answer your question about the advantage of the commercial tripwire above the open source one, but I use aide for some time now and it does a good job for me. I think it does the same as tripwire does (the version of 2 years ago, since it was the last time I used this).
. /etc/aide/aide.conf. See manual.html for information on this file.
[marc@scorpius marc]$ apt-cache show aide
Package: aide
Priority: optional
Section: admin
Installed-Size: 980
Maintainer: Mike Markley
Architecture: i386
Version: 0.9-2
Depends: libc6 (>= 2.3.1-1), debconf (>= 0.2.0)
Recommends: cron, mailx
Filename: pool/main/a/aide/aide_0.9-2_i386.deb
Size: 346316
MD5sum: a3610146e79608a34997450fdc56d74f
Description: Advanced Intrusion Detection Environment
AIDE creates a database from the regular expression rules that it finds
from the config file. Once this database is initialized it can be used to
verify the integrity of the files. It has several message digest algorithms
(md5,sha1,rmd160,tiger,haval,etc.) that are used to check the integrity of
the file. More algorithms can be added with relative ease. All of the usual
file attributes can also be checked for inconsistencies.
You will almost certainly want to tweak the configuration file in
Genius doesn't work on an assembly line basis. You can't simply say, "Today I will be brilliant."
About 4 months ago, a Windows-knowledgable colleague and my Unix-using self did a comparison of TripWire, ViperDB, Aide, Fcheck and another tool whose name escapes me. We were looking for speed, simplicity, effectiveness and portability *nix/Win32).
....". A little bit of scripting will do everything else for you.
FCheck ruled the day. It's easy to configure, works on *nix and Win32 (it's written in Perl), very fast in operation (We found Tripwire to be unusably slow/CPU-intensive for regularly scheduled checks) and passed every functional test we threw at it. It logs to syslog so you can send output to a remote machine. And it's GPL'd.
As for Tripwire's proprietary version, my colleague reckoned the only benefit was the GUI. Personally I don't see the point of a GUI on a security tool which is meant to run unsupervised. I suppose it does reporting etc. but really, what more do you need other than "This file changed at dd/mm/yy, hh:mm.ss. The change was
Yours Sincerely, Michael.
First off, if you only have to worry about a couple of machines, anything works pretty well.
Tripwire is good because it uses multiple hashing routines to figure out if something has changed (ie you can't pad a file with "0" until the hash is the same).
Additionally, the real strength in the commercial version of tripwire is the scalability. If you have hundreds of machines you need to monitor, the commercial version provides a central console which at a glance shows you what's going on across all your machines as far as changes. And the central console allows you to reconcile changes or revert to a known good state remotely.
All in all, if you only have a few boxen, don't buy it. If you have many and you don't want to spend all your time reconfiguring and updating a rules database, go for the commercial version.
I also was looking to use Tripwire mainly to occassionally scan the system to ensure that no important files had been modified (duh). I was extremely put off by the price and tone of the website.
If your main interest is simply to retain a database with checksums of files on your drive, and occassionally compare them for new files/changes - roll you own. I did and it was both easy an effective.
Simply stated, I use a configuration file to specify what directories and/or files should be scanned. Likewise, the configuration file has filters that will reject scanning files if any part of the filename matches the filter. The program reads the config and then goes out and reads the files on the drive. I use two different checksum schemes that produce checksum strings of about 80 characters each. These are stored in a database with the absolute file name, it's inode, it's last modify date, it's size, and the checksums.
When the program scans it merely checks the files against the database. If a file is new, it reports it as new to a log and adds it to the database. If a file has changed it reports it as changed to the log and then corrects the information in the database to reflect the change. If no change has occured, nothing happens to the database.
The program spits out little run-time facts about how many files it's scanning, number new, number change and number unchanged. When the run is completed all you have to do is glance at the log and determine if any of the files that changed in the log are a concern and need to be checked out.
There are a couple of advantages to do it yourself... first, no fee to Tripwire. Second... Tripwire is a known product. If you get a hacker in your system and he finds tripwire you can bet he'll try to do something to circumvent it. On the other hand, having written your own tripwire (and don't call it tripwire) - the hacker will not know this, not be familiar with your mechanism, and thus, will be unable to circumvent it. And finally, if your scanner is pretty good, clean and useable, it becomes a nice competitive product against Tripwire.
I never understood the requirement to have central management consoles for everything you run.
If you have so many servers that managing them individually is not an option, then what you need is a general solution to the management problem, not a specific solution for every piece of software you run.
For command line tools, manymaint (a nice Expect script) is one simple and free solution.
As for doing checks of routers, you could just use tftp to download configs to a server on a scheduled basis and run your checks there.
Computing is fun when you use your imagination to solve a problem (even an easy one like this) creatively, instead of asking "Here's my niche problem, where is the expensive niche product from a faceless bland corporation that fixes it?".
Yours Sincerely, Michael.
I spent a few weeks checking out various opensource and commercial packages. If you have less then 10 or 20 machines then you can use almost anything including the academic source release of tripwire. If you have more then 20 machines none of the opensource products that I found support centralized management/reporting/logging which is key to a large number or systems. Tripwire has a great product commercially wise but they are very expensive. I highly suggest you check out INTACT from pedestal software instead of tripwire. They are a third the price and have all the same functionality of tripwire and then some... I demo'd them for quite a while and it works very well on solaris/linux/windows. I dont have any relationship to them.. I was just impressed with their product.