I guess I have a tendency to write about negative articles more than about good ones. Lemme try to counterbalance this. A couple of years ago I met a gentleman called John McCumber at a DHS workshop. Very nice guy. We had some good discussions. Shortly after meeting him, I realized that he is a frequent author of articles in security magazines. I keep reading his articles and I have to say, I like them. They are generally very entertaining. He usually talks about something that happend to him in his daily life and translates that to the security world. If you have a chance read one of his articles.

This guy lists 12 steps in his article about how to approach a security program. I did not read all the twelve steps, but I found the one that’s of interest to me: “Step 8: Log reporting”. I started reading the paragraph and well, you bet, I have some comments:

“Management should know where the users are going, what type of bandwidth is being used, and who is hacking into your sites.”

Do you really believe that management is interested in where every user is going and what type of bandwidth they are using? I think they have better things to do. What about deifning a policy that clearly states what employees are allowed to do, what sites are off limit and what applications are prohibited (such as file sharing)? Then you monitor the traffic and figure out who is in violation of this. That’s the report that I as a manager would be interested in. I don’t have the time to interpret log files or reports and figure out what happened. Have machines do the work for me and give me the distilled information!

A horrible acronym. I know. We had a working session during the RSA conference to talk about XCCDF-P. For those not familiar with XCCDF, it has to do with policy definitions and uses OVAL to implement the checks.

XCCDF-P (which will hopefully get renamed pretty soon to something else, and hopefully not to CPN (Common Platform Names) [We already have CVE, CME, and CCE]) is an effort to standardize platform names. What’s the problem? Well, if I have two scanners analyzing a system of mine, one of them might report that I am running a “Windows 2000”, the other one might say “Win2K”. This is really the same, but how would a machine know? That’s where the standard is trying to clean things up. You wouldn’t belive how much discussion this topic actually involves. We met for about an hour and had plenty of things to discuss, not even closely getting to an agreed-upon solution. However, the problem is defined and we all agreed upon the the necessity to solve the problem! Stay put for an update soon and hopefully a quick turn around with a solution draft.

I am frustrated. I found these Perl libraries to buld treemaps (Treemap::Squarified). The problem is that it needs this special input format, which is either an XML tree or you can hack it into the internal perl datastructure which are basically a bunch of arrays. Once I figured out the internal datastructure (I was too lazy to go through XML), it got worse. You need to do everything yourself. The library does not even take care of sizing the pieces for you. You need to make sure that the numbers along the hierarchy are all correct and add up. But that’s not all. After playing with that for a while (basically my problem was to convert CSV to a tree; no I did not finish implementing it), I got into GTK2 coding. Well, that’s a mess too. Hardly anything documented. I just wanted to show some pictures in a window. Easy? No! I wanted to resize them to fit two arbitrary images into one window. Resize? I could not quite figure out how to have pixbufs and Gtk2 and all that interact. So I gave up…
Back to a language that I know a bit better: JAVA. Starting over…

I just read this article about ITM – yet another acronym, which stands for insider threat management. Looking at the products they reviewed in the infoworld article, I just don’t see what is so new about them. They seem to be either a NIDS or a HIDS on steroids. Why wouldn’t I be able to flag specific traffic with a NIDS? I can build a rule in snort which looks for SSN numbers floating around on my network. Yes, there are some nice managebility features built into these new products, but why don’t NIDS vendors add them on top of their products? I am over-simplifying, but think about it, all these new products are not really _that_ new. They wrap old concepts in new products.
Is marketing taking over completely?

Wow. I just found this pretty awesome TreeMap tool. The data format it reads is pretty easy and I quickly built a file with some of my firewall data. Well, fake firewall data 😉

What you see in here is first the color: green are firewall passes, red are blocks, then the hierarchy is such that the target system is top, then the target IP and then you see the date inside of the boxes, when the access happened.

Well, the tool is pretty awesome. Lots of interactivity. You define the hierarchies manually, on the fly it updates the graph. Then you can color and filter and all kinds of nifty things. Try it out.

It’s one of those afternoons again where I am loosing myself in doing some research on graph tools. Maybe I can spare some others an afternoon of browsing around and summarize some things I found:

• kgraphviewer – this is a tool to show graphics. The exciting thing is that you can open .dot files. However, only dot is used to render them, not neato or twopi. Too bad.
• ZGRViewer – looks very promising. You can open .dot files and render them with anything (neato, twopi) the output is zoomable and you can move around. Pretty nice. It’s written in Java. Now the question is whether I could use this framework to build my animated viewer that reads the next .dot file and updates the graph nicely. I will get it done someday!
• tinfu – Another java library that draw .dot files. It had trouble with some of my .dot files. I guess the tool’s not very robust.
• Grappa – Another Java tool. The feature list looks promising. The Web demo is interesting. Need to download it and have a closer look.

I guess one thing that I realized is that you could principally build a graph with, let’s say twopi, and generate an SVG. Then you can manipulate that, show it to the user and so on. The question is still how to periodically update a graph without totally re-layouting it. Well, I am working on this. I think I just haven’t really understood the problem yet 😉

While playing with the AT&T graphviz library again, I was trying to get nodes to show up as pictures, which is fairly simple:

node [peripheries=0, fontsize=10, shapefile="computer.gif"]

However, what I really want is a transparent gif (computer.gif) and then using background color to color that node. That way I could programmatically change the color of the nodes and still use icons to make the nodes a bit nicer. This is what I tried:

node [peripheries=1, fontsize=10, shapefile="computer.gif", fillcolor=blue]

Unfortunately it did not work. Anyone?

Sitting down at a cafe around the corner from where I live, I realize that some of the scripts I wrote a while back might actually benefit others too. This one is to connect to the first available access point:

#!/bin/bash

iwconfig ath0 ap $ap
iwconfig ath0 essid $essid
iwconfig ath0 nick test
killall -9 dhclient
dhclient ath0


Not sure whether there would be a simpler solution natively supported by linux…

I haven’t looked at Shoki in a while. Today I downloaded a version again and tried to compile it on my Fedora Core 4 installation, just to find out that the thing would not compile. Well, I dug around in the code for a bit and after some searches on the Web, I realized that gcc 4 is stricter about the C conventions and Shoki was written with some declarations being non-standard. What fixed it was to define the CC flag in the Makefile to use gcc32 instead of gcc.
Playing with this tool, I somehow have the impression that I just don’t get it. I can redefine the axes and play with that, but even zomming into a certain selection I can’t seem to accomplish. And then there is all this extra stuff like fast fourier transformations etc. While I know what that is, I just don’t quite understand how all that works in Shoki. Maybe I have to spend an afternoon with the documentation 😉 Or maybe there are people out there who have some tips or hints for me?
What I am really interested in is if someone managed to analyze a dataset and can show me what he found with what feature. Do all the bells and whistles (some of the advanced features) really help? Help me out!