Sunday, November 02, 2014

toolsmith: Inside and Outside the Wire with FruityWifi & WUDS

I recommend a dedicated (non-VM) Kali distribution if you don’t have a Raspberry Pi.

I have noted to myself, on more than one occasion, now more than eight years in to writing toolsmith, that I have not once covered wireless assessment tools. That constitutes a serious shortcoming on my part, one that I will rectify here with a discussion of FruityWifi and WUDS. These tools serve rather different purposes but both conform to the same principle of significant portability as both run on Raspberry Pi. Both also run on Debian systems (Kali) which is how I ran both for toolsmith testing purposes. FruityWifi is an open source platform with which to audit wireless networks, allowing users to conduct various attacks via the web interface or remote messaging. It is modular, feature-rich, and just celebrated a v2.0 release with many upgrades. WUDS, or the Wi-Fi User Detection System, on the other hand, is a proximity-based physical security concept that alerts on unapproved Wi-Fi probe requests bouncing off a WUDS sensor. Per the WUDS introduction, “The combination of a white list of unique identifiers for devices that belong in the area (MAC addresses) and signal strength (RSSI) can be used to create a protected zone. With tuning, this creates a circular detection barrier that, when crossed, can trigger any number of alert systems. WUDS includes an SMS alert module, but the sky is the limit.”
WUDS comes to us courtesy of toolsmith alum Tim Tomes (@lanmaster53) whose Recon-ng, the 2013 Toolsmith Tool of the Year, we covered in May 2013.

For FruityWifi highlights I reached out to xtr4nge (@FruityWifi), the project lead/developer.
The initial idea was to create an open source application to audit wireless networks and perform penetration tests from a Raspberry-Pi, or any other platform or device in a flexible, modular and portable manner.
Soon after the first version was published, FruityWifi was presented to a Rooted Warfare Spain audience (Rooted CON) in March 2014. FruityWifi was well received at the conference and by users from the onset, and many users sent feedback and ideas to improve it. 
A new version of FruityWifi (v2.0) was published a few weeks ago featuring many changes and updates; a new interface, new modules, Realtek chipsets support (confirmed in my testing with my Alfa card), Mobile Broadband (3G/4G) support, a new control panel, and more. The tool is under constant development and new modules and improvements are being published regularly.

Tim kindly provided detail regarding his favorite WUDS features and use case. His favorite feature is the alert system, and I strongly second this; I’ll show you why later in the article. Tim built a really simple interface for creating new alerts for the system, which does not require the need to dig into core components of the code to create new alerts. You need only add a function to the file, name it correctly, and add associated options to the file, and you’re finished. Tim’s script originally included just an SMS alerting mechanism, but he has since added a Pushover notification alert (5 minutes and 8 lines of code to implement); he exclusively uses the Pushover alert.
Tim’s favorite WUDS use case story is cited in his article, about the deliveryman alerting the system during testing. He’d been doing some testing the night before and in the middle of the afternoon the following day, an SMS message came through to his phone notifying him that someone had crossed his detection barrier. He was about to write it off as a false positive when the doorbell rang and it was a delivery service dropping off a package. The alert served as positive affirmation that the concept is a sound one.
In the future, Tim plans to either expand WUDS, or create another tool all together, that does the exact opposite of WUDS. Rather than alert on foreign MAC addresses, this tool would allow the user to configure the sensor to alert when certain MAC addresses leave the premises during specified windows of time. This would provide a sort of latchkey system that is not challenged by MAC randomization issues; devices will be properly connected to the local WAP with their normal MAC addresses when on premises. That said, the tool would need to be expanded to sense more than just probes.
I ran FruityWifi and WUDS on a dedicated Lenovo T61p laptop with Kali 64-bit installed and utilized the onboard wireless adapter.  Both tools are optimized to perform well on Raspberry Pi, but as I don’t have one, I experimented with both Ubuntu and Kali and was far more satisfied with Kali, no muss, no fuss. All installation steps that follow assume you’re running on Kali.

FruityWifi installation

FruityWifi installation is very simple. Download the master zip file or git clone the repository to your preferred directory, then cd /FruityWifi from there. Run ./ If you have any issues after installation where FruityWifi isn’t available via the browser, it may be related to the Nginx/PHP5-FPM deployment. You can follow the FruityWifi Nginx wiki guidance to correct the issue. Thereafter, browse to http://localhost:8000 or https://localhost:8443, login with admin and admin (change the password), and you’re off to the races as seen in the UI’s configuration page per Figure 1.

Figure 1 – FruityWifi configuration page
FruityWifi inside the perimeter

Building on the same principles as the Pwn Plug, a FruityWifi-enable device can wreak havoc once unleashed inside any given network. There are a significant number of modules you can install and enable, a veritable fruit basket, depending on what you wish to accomplish, as seen in Figure 2.

Figure 2 – FruityWifi modules galore
If you utilized the earlier version of Fruity, you’ll really appreciate the update that is 2.0. Clean, fast, intuitive, and lots of fresh functionality. Red-teamers will enjoy AutoSSH, which allows reverse SSH connections, and automatic restart for connections that have been closed or dropped. FruityWifi 2.0 includes Nessus, Nmap, and Meterpreter as well. MDK3 is particularly attractive if you’re conducting an aggressive pentest and you want to create a distraction or a disruption.  You had better have permission before going off with MDK3, wireless hacking is deemed criminal in more than one state. MDK, or murder, death, kill for WLAN environments, utilizes a variety of SSID, authentication, and de-authentication flooding techniques to create wireless DoS conditions, and on occasion, WLAN hardware resets. My favorite recent addition to FruityWifi isn’t one of the hacking or enumeration tools, it’s actually vFeed from our friend @toolswatch. To quote vFeed’s description from the Fruity UI, is a vulnerability database (SQLite) that “provides extra structured detailed third-party references and technical characteristics for a CVE entry through an extensible XML schema.” You can search it right on your FruityWifi instance, after you’ve run Nmap and Nessus scans, identified potentially vulnerable targets and want to look up the related CVE. The available data includes:
  • Open security standards: CVE, CWE, CPE, OVAL, CAPEC (all per Mitre), and CVSS
  • Vulnerability Assessment & Exploitation IDs: Metasploit, Saint Corporation, Nessus Scripts, Nmap, Exploit-DB, milw0rm
  • Vendors Security Alerts: Microsoft, Debian, Redhat, Ubuntu, and others
I looked up CVE-2013-3893, as seen in Figure 3, and was treated to a summary and exploit details. Take note of the vFeed export feature as well. I love vFeed so much I wrote an R parser to turn the XML export into human readable Excel docs for broad reporting and consumption without the machine layer. I’ll be sharing that via the HolisticInfoSec blog and website.

Figure 3 – FruityWifi’s vFeed module informs the analyst
FruityWifi represents a fabulous way to establish a foothold inside a given perimeter, pivot to additional targets, and conduct complete compromise. Let’s now explore WUDS, intended to help you defend the perimeter. First a little red, then a little blue. Wi-fi not?

WUDS Installation

Tim’s Bitbucket installation guidance is short and sweet:
sudo apt-get install iw python-pcapy sqlite3 screen
# launch a screen session
# install WUDS
git clone
cd wuds
# edit the config file
# execute the included run script
You really need to get your implementation correct. Default settings work well initially until you get to your ALERT_SMS CONFIG. You’ll need SMTP server access, including the outgoing SMTP server with the TLS port along with username and password, in order to send alert messages. Android and iOS users (there are browser plugins too) can also take advantage of a Pushover account, as Tim mentioned.
Debug is enabled by default, if there are issues, when you run ./ you’ll receive failure notice.

WUDS defends the perimeter

Once WUDS is running there’s not a whole lot to actually see. No sexy UI, just a SQLite database and alerts of your choosing. Figure 4 represents a sqlite browser view of logs.db the WUDS datastore.

Figure 4 – A view to the WUDS database
You’ll note that the detected devices all have received signal strength indications (RSSI) of higher than -50. Recall, how much I stressed The default RSSI threshold for triggering alerts is -50, but you can adjust it depending on how you wish to define your perimeter.  
The real pleasure comes from the first alerts received on your mobile device. As you see in a screen shot from my phone (Figure 5), proximity alerts advise me that a variety of devices have been detected on the premises. Ruh-roh!

Figure 5 –WUDS alerts of perimeter violations
True story. When I first enabled WUDS in my office at work, I immediately received alerts for an HP device that was beaconing for my home wireless AP. This freaked me out for a minute as I knew of no HP devices currently in use and certainly not those looking for my house infrastructure. After looking around again, and calming down a bit, a spotted it, an old HP printer under my desk that I’d brought in for scanning, had turned it on years ago, and literally forgotten about it ever since. And there it was, blindly beaconing away for a WAP it would never again communicate with. Thanks WUDS!
You’ll find all sorts of interesting devices chattering away when you enable WUDS. Just remember, the more dense the population area, the noisier it will be. Avoid self-induced mobile device DoS attacks. :-)

In Conclusion

Great tools from xtr4nge and Tim, I’m thrilled to have gotten off the schnide regarding wireless topics with FruityWifi and WUDS. I’m thinking there’s actually an opportunity to incorporate WUDS in FruityWifi. You heard it here first. Enjoy these tools, they’re both a ton of fun and incredibly useful at the same time.
Ping me via email if you have questions (russ at holisticinfosec dot org).
Cheers…until next month.


xtr4nge, FruityWifi project lead and developer
Tim Tomes, WUDS project lead and developer

Friday, October 03, 2014

toolsmith: HoneyDrive - Honeypots in a Box


Virtualization platform


Late in July, Ioannis Koniaris of BruteForce Lab (Greece) released HoneyDrive 3, the Royal Jelly edition. When Team Cymru’s Steve Santorelli sent out news of same to the Dragon News Bytes list the little light bulb went off in my head. As I prepared to write our ninety-sixth toolsmith for October’s edition I realized I had not once covered any honeypot technology as the primary subject matter for the monthly column. Time to rectify that shortcoming, and thanks to Ioannis (and Steve for the ping on DNB radar screen) we have the perfect muse in HoneyDrive 3.
From HoneyDrive 3’s own description, it’s a honeypot Linux distro released as a virtual appliance (OVA) running Xubuntu Desktop 12.04.4 LTS edition which includes over 10 pre-installed and pre-configured honeypot software packages. These includes the Kippo SSH honeypot, Dionaea and Amun malware honeypots, the Honeyd low-interaction honeypot, Glastopf web honeypot and Wordpot, Conpot SCADA/ICS honeypot, as well as Thug and PhoneyC honeyclients and more. It also includes many useful pre-configured scripts and utilities to analyze, visualize and process the data it captures, such as Kippo-Graph, Honeyd-Viz, DionaeaFR, an ELK stack and much more. Finally, nearly 90 well-known malware analysis, forensics and network monitoring related tools are included with HoneyDrive 3   .
Ioannis let me know he started HoneyDrive mostly out of frustration arising from the difficulty of installing and configuring some of the well-known honeypot systems. At first, he created scripts of his own to automate their installation and deployment but then decided to put them all in a nice package for two reasons:
1.       For newcomers to be able to quickly deploy and try out various honeypot systems,
2.       To connect the honeypot software with all the existing projects built on top of them.
As an example Ioannis developed Kippo-Graph, Honeyd-Viz and various other tools while HoneyDrive makes the integration between the backend (honeypots) and frontend (tools) seamless. Ioannis has strong evidence that HoneyDrive and some of the specific tools he’s created are very popular based on the interactions he’s had online and in-person with various researchers. HoneyDrive is used in many universities, technical research centers, government CERTs, and security companies. Ioannis believes honeypots are more relevant than ever given the current state of global Internet attacks and he hopes HoneyDrive facilitates their deployment. His roadmap includes creating visualization tools for honeypot systems that currently don't have any visualization features, and attempt to develop a way to automatically setup HoneyDrive sensors in a distributed fashion.
This is a great effort, and it really does not only simplify setup and getting underway, but the visual feedback is rich. It’s like having a full honeypot monitoring console and very easy to imagine HoneyDrive views on big monitors in security operations centers (SOC). Ready to give it a try?

HoneyDrive Prep

Download the HoneyDrive OVA via SourceForge. This is a fully configured 4GB open virtual appliance that you can import into your preferred virtualization platform. I did so on VMWare Workstation 10, which complained a bit initially but gave me the option to bypass its whining and proceed unfettered. There’s a good convert-to-VMWare doc if you need it but I conducted a direct import successfully. Royal Jelly has run like a champ since. If you’re exposing the virtual machine in order to catch some dirty little flies in your honey traps keep in mind that your virtual network settings matter here. Best to bridge the VM directly to the network on which you’re exposing your enticing offerings, NAT won’t work so well, obviously. Apply all the precautions associated with hosting virtual machines that are likely to be hammered. Depending on where you deploy HoneyDrive and the specific honeypots you plan to utilize, recognize that it will be hammered, particularly if Internet facing. Worn out, rode hard and put away wet, flogged…hammered. Feel me? The beauty is that HoneyDrive does such a fabulous job allowing for performance monitoring, you’ll be able to keep an eye on it. With virtualization you can always flush it and restart from your snapshot, just remember to ship off your logs or databases so you don’t lose valuable data you may have been collecting. Let’s play.

I Am Honeydripper, Hear Me Buzz

There is SO much fun to be had here, where to begin? Rhetorical…we begin with carefully reading the comprehensive README.txt file conveniently found on the HoneyDrive desktop. This README describes all available honeypots and their configurations. You’ll also find reference to the front-end visualization offerings such as Ioannis’ Kippo-Graph. Perfect place to get started, Kippo is a favorite.


Kippo, like all its counterparts found on HoneyDrive, is available as a standalone offering, but is ready in an instant on HoneyDrive. From a Terminator console, cd /honeydrive/kippo followed by ./ You should receive Starting kippo in the background...Loading dblog engine: mysql. You’re good to go. If you need to stop Kippo it’s as easy as…wait for it,./ From a remote system attempt an SSH connection to your HoneyDrive IP address and you should meet with success. I quickly fired up my Kali VM and pounded the SSH “service” the same way any ol’ script kiddie would: with a loud bruteforcer. My favorite it is Patator using the SSH module and the little John dictionary file from fuzzdb as seen in Figure 1.

Figure 1: Bruteforcing Kippo’s SSH service with Patator
As you can see my very first hit was successful using that particular dictionary. Any knucklehead with 123456 in their password lists would think they’d hit pay dirt and immediately proceed to interact. Here’s where Kippo-Graph really shines. Kippo-Graph includes visual representations of all Kippo activity including Top 10s for passwords, usernames, combos used, and SSH clients, as well as success ratios, successes per day/week, connections per IP, successful logins from same IP, and probes per day/week. Way too many pretty graphs to print them all here, but Kippo-Graph even includes a graph gallery as seen in Figure 2.

Figure 2: Kippo-Graph’s Graph Gallery shines
But wait, there’s more. I mentioned that a bruteforce scanner who believes they are successful will definitely attempt to login and interact with what they believe is a victim system. Can we track that behavior as well? Yah, you betcha. Check out Kippo-Input, you’ll see all commands passed by attackers caught in the honeypot. Kippo-Playlog will actually playback the attacker’s efforts as video and offering DIG and location details on the related attacker IP. Figure 3 represents Kippo-Input results.

Figure 3: Kippo-Input provides the attacker’s commands
Many of these graphs and visualizations also offer CSV output; if you wish to review data in Excel or R it’s extremely useful. HoneyDrive’s Kippo implementation also allows you to store and review results via the ELK (Elasticsearch, Logstash, Kibana) stack, using Kippo2ElasticSearch, that we first introduced in our toolsmith C3CM discussions.
Of course, Kippo is not the only honeypot offering on HoneyDrive 3, let’s explore further.


Per the Honeynet Project site, “Dionaea is a low-interaction honeypot that captures attack payloads and malware. Dionaea is meant to be a nepenthes successor, embedding python as scripting language, using libemu to detect shellcodes, supporting ipv6 and tls.”
HoneyDrive includes the DionaeaFR script which provides a web UI for all the mayhem Dionaea will collect.
To start Dionaea, first cd /honeydrive/dionaea-vagrant then run ./ Follow this with the following to start DionaeaFR:
cd /honeydrive/DionaeaFR/
python manage.collectstatic
python runserver
Point your browser to http://[your HoneyDrive server]:8000 and you’ll be presented a lovely UI Dionaea.
Even just an NMAP scan will collect results in DionaeaFR but you can also follow Dionaea with Metasploit to emulate malware behavior. Figure 4 is a snapshot of the DionaeaFR dashboard.

Figure 4: DionaeaFR dashboard
You can see connection indicators from my NMAP scan as well as SMB and SIP exploits attempts as described in Emil’s Edgis Security blog post.


Wordpot is a WordPress honeypot. No one ever attacks WordPress, right? Want to see how badly WordPress is attacked en masse when exposed to the Internet? Do this:
sudo service apache2 stop (WordPot and Apache will fight for port 80, suggest moving Apache to a different port anyway)
cd /honeydrive/wordpot
sudo python
You’ll find the logs in /honeydrive/Wordpot/logs. My logs, as represented along with my fake WordPress site in Figure 5, are the result of a Burp Suite scan I ran against it. If you expose WordPot to the evil intarwebs, your logs will look ridiculously polluted by comparison.

[Insert wordpot.png]
Figure 5: WordPot site and WordPot logs

A number of HoneyDrive offerings write to SQLite databases. Lucky for you, HoneyDrive includes phpLiteAdmin, a web-based SQLite database admin tool (like phpMyAdmin). Note that is configured to accept traffic only from localhost by default.

In Conclusion

This is such a great distribution, I’m thrilled Ioannis’ HoneyDrive is getting the use and attention it deserves. If you haven’t experimented or deployed honeypots before you quite literally have no excuse at this point.  As always, practice safe honeypotting, no need to actually suffer a compromise. Honeypots need to be closely monitored, but that’s exactly what makes HoneyDrive so compelling, great visualization, great logging, and great database management. HoneyDrive is certainly a front runner for toolsmith tool of the year, but that, as always, is up to you, my good reader. Download HoneyDrive ASAP and send me feedback.
Ping me via email if you have questions (russ at holisticinfosec dot org).
Cheers…until next month.


Ioannis Koniaris, project lead and developer

Monday, September 01, 2014

toolsmith - Jay and Bob Strike Back: Data-Driven Security


Data-Driven Security: Analysis, Visualization and Dashboards
R and RStudio as we’ll only focus on the R side of the discussion
All other dependencies for full interactive use of the book’s content are found in Tools You Will Need in the books Introduction.

When last I referred you to a book as a tool we discussed TJ O’Connor’s Violent Python. I’ve since been knee deep in learning R and quickly discovered Data-Driven Security: Analysis, Visualization and Dashboards from Jay Jacobs and Bob Rudis, hereafter referred to a Jay and Bob (no, not these guys).

Jay and Silent Bob Strike Back :-)
Just so you know whose company you’re actually keeping here Jay is a coauthor of Verizon Data Breach Investigation Reports and Bob Rudis was named one of the Top 25 Influencers in Information Security by Tripwire.
I was looking to make quick use of R as specific to my threat intelligence & engineering practice as it so capably helps make sense of excessive and oft confusing data. I will not torment you with another flagrant misuse of big data vendor marketing spew; yes, data is big, we get it, enough already. Thank goodness, the Internet of Things (IoT) is now the most abused, overhyped FUD-fest term. Yet, the reality is, when dealing with a lot of data, tools such as R and Python are indispensable particularly when trying to quantify the data and make sense of it. Most of you are likely familiar with Python but if you haven’t heard of R, it’s a scripting language for statistical data manipulation and analysis. There are a number of excellent books on R, but nowhere will you find a useful blending of R and Python to directly support your information security analysis practice as seen in Jay and Bob’s book. I pinged Jay and Bob for their perspective and Bob provided optimally:
“Believe it or not, we (and our readers) actually have ZeroAccess to thank for the existence of Data-Driven Security (the book, blog and podcast). We started collaborating on security data analysis & visualization projects just about a year before we began writing the book, and one of the more engaging efforts was when we went from a boatload of ZeroAccess latitude & longitude pairs (and only those pairs) to maps, statistics and even graph analyses. We kept getting feedback (both from observation and direct interaction) that there was a real lack of practical data analysis & visualization materials out there for security practitioners and the domain-specific, vendor-provided tools were and are still quite lacking. It was our hope that we could help significantly enhance the capabilities and effectiveness of organizations by producing a security-centric guide to using modern, vendor-agnostic tools for analytics, a basic introduction to statistics and machine learning, the science behind effective visual communications and a look at how to build a great security data science team.
One area we discussed in the book, but is worth expanding on is how essential it is for information security professionals to get plugged-in to the broader "data science" community. Watching "breaker-oriented" RSS feeds/channels is great, but it's equally as important to see what other disciplines are successfully using to gain new insights into tough problems and regularly tap into the wealth of detailed advice on how to communicate your messages as effectively as possible. There's no need to reinvent the wheel or use yesterday's techniques when trying to stop tomorrow's threats.”
Well said, I’m a major advocate for the premise of moving threat intelligence beyond data brokering as Bob mentions. This books endeavors and provides the means with which to conduct security data science. According to Booz Allen’s The Field Guide to Data Science, “data science is a team sport.” While I’m biased, nowhere is that more true than the information security field. As you embark on the journey Data-Driven Security: Analysis, Visualization and Dashboards (referred to hereafter as DDSecBook) intends to take you on you’ll be provided with direction on all the tools you need, so we’ll not spend much time there and instead focus on the applied use of this rich content. I will be focusing solely on the R side of the discussion though as that is an area of heavy focus for me at present.  DDSecBook is described with the byline Uncover hidden patterns of data and respond with countermeasures. Awesome, let’s do just that.

Data-Driven Security

DDSecBook is laid out in such a manner as to allow even those with only basic coding or scripting (like me; I am the quintessential R script kiddie) to follow along and grow while reading and experimenting:
1.       The Journey to Data-Driven Security
2.       Building Your Analytics Toolbox: A Primer on Using R and Python for Security Analysis
3.       Learning the “Hello World” of Security Data Analysis
4.       Performing Exploratory Security Data Analysis
5.       From Maps to Regression
6.       Visualizing Security Data
7.       Learning from Security Breaches
8.       Breaking Up with Your Relational Database
9.       Demystifying Machine Learning
10.   Designing Effective Security Dashboards
11.   Building Interactive Security Visualizations
12.   Moving Toward Data-Driven Security

For demonstrative purposes of making quick use of the capabilities described, I’ll focus our attention on chapters 4 and 6. As a longtime visualization practitioner I nearly flipped out when I realized what I’d been missing in R, so chapters 4 and 6 struck close to home for me. DDSecBook includes code downloads for each chapter and the related data so you can and should play along as you read. Additionally, just to keep things timely and relevant, I’ll apply some of the techniques described in DDSecBook to current data of interest to me so you can see how repeatable and useful these methods really are.

Performing Exploratory Security Data Analysis

Before you make use of DDSecBook, if you’re unfamiliar with R, you should read An Introduction to R, Notes on R: A Programming Environment for DataAnalysis and Graphics and run through Appendix A. This will provide at least an inkling of the power at your fingertips.
This chapter introduces concepts specific to dissecting IP addresses including their representation, conversion to and from 32-bit integers, segmenting, grouping, and locating, all of which leads to augmenting IP address data with the likes of IANA data. This is invaluable when reviewing datasets such as the AlienVault reputation data, mentioned at length in Chapter 3, and available as updated hourly.
We’ll jump ahead here to Visualizing Your Firewall Data (Listing 4-16) as it provides a great example of taking methods described in the book and applying it immediately to your data. I’m going to set you up for instant success but you will have to work for it a bit. The script we’re about to discuss takes a number of dependencies created earlier in the chapter; I’ll meet them in the script for you (you can download it from my site), but only if you promise to buy this book and work though all prior exercises for yourself. Trust me, it’s well worth it. Here’s the primary snippet of the script, starting at line 293 after all the dependencies are met. What I’ve changed most importantly is the ability to measure an IP list against the very latest AlienVault reputation data. Note, I found a bit of a bug here that you’ll need to update per the DDSecBook blog. This is otherwise all taken directly ch04.r in the code download with specific attention to Listing 4-16 as seen in Figure 2.

FIGURE 2: R code to match bad IPs to AlienVault reputation data
I’ve color coded each section to give you a quick walk-through of what’s happening.
1)      Defines the URL from which to download the AlienVault reputation data and provides a specific destination to download it to.
2)      Reads in the AlienVault reputation data, creates a data frame from the data and provides appropriate column names. If you wanted to read the top of that data from the data frame, using head(av.df, 10) would result in Figure 3.

FIGURE 3: The top ten entries in the Alien Vault data frame
3)      Reads in the list of destination IP addresses, from a firewall log list as an example, and compares it against matches on the reliability column from the AlienVault reputation data.
4)      Reduces the dataset down to only matches for reliability above a rating of 6 as lower tends to be noise and of less value.
5)      Produces a graph with the function created earlier in the ch04.r code listing.
The results are seen in Figure 4 where I mapped against the Alien Vault reputation data provided with the chapter 4 download versus brand new AlienVault data as of 25 AUG 2014.

FIGURE 4: Bad IPs mapped against Alien Vault reputation data by type and country
What changed, you ask? The IP list provided with chapter 4 data is also a bit dated (over a year now) and has likely been cleaned up and is no longer of ill repute. When I ran a list 6100 IPs I had that were allegedly spammers, only two were identified as bad, one a scanning host, the other for malware distribution. 
Great stuff, right? You just made useful, visual sense of otherwise clunky data, in a manner that even a C-level executive could understand. :-)

Another example the follows the standard set in Chapter 6 comes directly from a project I’m currently working on. It matches the principles of said chapter as built from a quote from Colin Ware regarding information visualization:
“The human visual system is a pattern seeker of enormous power and subtlety. The eye and the visual cortex of the brain form a massively parallel processor that provides the highest bandwidth channel into human cognitive centers.”
Yeah, baby, plug me into the Matrix! Jay and Bob paraphrase Colin to describe the advantages of data visualization:
·         Data visualizations communicate complexity quickly.
·         Data visualizations enable recognition of latent patterns.
·         Data visualizations enable quality control on the data.
·         Data visualizations can serve as a muse.
To that end, our example.
I was originally receiving data for a particular pet peeve of mine (excessively permissive open SMB shares populated with sensitive data) in the form of a single Excel workbook with data for specific dates created as individual worksheets (tabs). My original solution was to save each worksheet as individual CSVs then use the read.csv function to parse each CSV individually for R visualization. Highly inefficient given the like of the XLConnect library that allows you to process the workbook and its individual worksheets without manipulating the source file.
raw <- data="" harestats0727.csv="" openshares="" read.csv="" span="">
h <- ostct="" raw="" span="" sum="">
s <- harect="" raw="" span="" sum="">
sharestats <- data="" harestats_8_21.xlsx="" loadworkbook="" openshares="" span="">
sheet1 <- readworksheet="" sharestats="" sheet="1)</span">
h1 <- ostct="" sheet1="" span="" sum="">
s1 <- harect="" sheet1="" span="" sum="">
The first column of the data represented the number of hosts with open shares specific to a business unit, the second column represented the number of shares specific to that same host. I was interested in using R to capture a total number of hosts with open shares and the total number of open shares over all and visualize in order to show trending over time. I can’t share the source data with you as its proprietary, but I’ve hosted the R code for you. You’ll need to set your own working directory and the name and the path of the workbook you’d like to load. You’ll also need to define variables based on your column names. The result of my effort is seen in Figure 5.

FIGURE 5: Open shares host and shares counts trending over time
As you can see, I clearly have a trending problem, up versus down is not good in this scenario.
While this is a simple example given my terrible noob R skills, there is a vast green field of opportunity using R and Python to manipulate data in such fashion. I can’t insist enough that you give it a try.

In Conclusion

Don’t be intimidated by what you see in the way of code while reading DDSecBook. Grab R and R Studio, download the sample sets, open the book and play along while you read. I also grabbed three other R books to help me learn including The R Cookbook by Paul Teeter, R for Everyone by Jared Lander, and The Art of R Programming by Normal Matloff. There are of course many others to choose from. Force yourself out of your comfort zone if you’re not a programmer, add R to your list if you are, and above all else, as a security practitioner make immediate use of the techniques, tactics, and procedures inherent to Jay and Bob’s most excellent Data-Driven Security: Analysis, Visualization and Dashboards.
Ping me via email if you have questions (russ at holisticinfosec dot org).
Cheers…until next month.


Bob Rudis, @hrbrmstr, DDSecBook co-author, for his contributions to this content and the quick bug fix, and Jay Jacobs, @jayjacobs, DDSecBook co-author.