The NSRL For Hackers

A short article I wrote for 2600 - The Hacker Quarterly. It is about the National Software Reference Library and what they and you can do with it.

2600 is going through tough times right now and I would appreciate it very much if you would buy the magazine,

The NSRL For Hackers

First published in 2600, 37:3

The National Software Reference Library (NSRL) is something you probably never heard of, although it is one of the largest software collections in the world. It is maintained by the National Institute of Standards and Technology (NIST), supported and used by Homeland Security, FBI, and other security agencies in the U.S. and the world. In this article, I will give you an overview over the NSRL and an idea about what you can do with it. And what they use it for.

0x0 What Is It And What’s In It

Firstly, the NSRL is a huge corpus of software: applications, operating systems, games, libraries, and tools. It holds proprietary products like Microsoft Windows or Adobe suites, but also open source things like Linux distributions or GNU compilers. The NSRL does not contain forbidden user data like terror propaganda videos or images of sexual abuse. Secondly, it is a huge collection of generated metadata sets from the products in the corpus. And these metadata sets are available for free and without registration. The metadata sets are called Reference Data Set Hash Sets (RDS hash sets) and there are six of them, each different.

0x1 RDS Types and Structure

All of these six sets consist of four simple - but huge - csv text files. In them, you find SHA-1, MD5, and CRC32 hash sums, file sizes, names, and more. Let’s have a closer look. The RDS Modern set has information about software from 2000 and later and counts over 104 million entries, with doublets. What is a doublet in this context? A wallpaper, for example, can be part of two operating system releases. The wallpapers are identical, i.e., have the same hash sum. As it is used in two places, the entry in the set is there twice and, therefore, the information is redundant. The RDS Modern Minimal is also about software from 2000 and later, but with doublets eliminated and has still more than 26 million entries. The RDS Modern Unique is again about software from 2000 and later, but consists only of entries that have no doublets in the first place. I have no clue why someone may use this set. RDS Legacy has more than 107 million entries and includes all the code before 2000.

Finally, we have two sets for mobiles, one each for iOS with more than 14 million and Android with 13 million entries. What do the entries look like, you ask? Here we go. Two examples, with the field names in the first line. These entries are from the NSRLFile.txt:



”Batman \_Seventies.POR”,90,196184,”362”,””


The first five fields are self-explanatory. ProductCode references to an entry in the file NSRLProd.txt. This file has more information on the products where the files are used. Let’s have a look into this file and search for the product code 17066:

17066,”Linux Mint 17.2 Rafaela Cinnamon 32-bit”,”2006”,”51”,”534”,”English”,”Operating System”

OK, so we know that the file bpa10x.ko is used in the operating system Linux Mint 17.2, 32-bit version. The field OpSystemCode references an operating system in the file NSRLOS.txt:


As you can see, we have relations between the information in the different metadata files of the set.

0x2 Usage Stories

Now it is nineteen eighty-four Knock-knock at your front door
It’s the suede denim secret police
They have come for your ... PC and hard
disks and USB sticks and mobiles and ...

Dead Kennedys in 1979

When some nosy people get their hands on your equipment and stored data, they are only interested in specific data. Data that is user- generated and therefore not in off-the-shelf products like operating systems. When those people have to check millions of files on your disks to find a specific file or information, it is handy to check if a file is in the NSRL. If it is not, chances are high that the file is worth a look. So hiding information in files, disguised as drivers in system folders, isn’t a good idea anymore. To be honest, it never was.

Since you’ve read this far, you’ll probably find all this interesting. But you might wonder why the NSRL should be of interest for your daily practical ITsec work. Let me give you two examples.

  1. Baseline for Intrusion Detection Systems You could use the NSRL as a baseline for an intrusion detection system. Extract all entries relevant for your operating system and compare the hashes.

  2. File carver If you have to restore files from a crashed hard disk, you could find yourself in the same situation like the law enforcement guys. When a disk is damaged, in most cases you can create an image with “dd” or something else. You then let a file carver like “scalpel” do its work on the image and carve files. If the file names or metadata cannot be restored, you could compare the hashes of the carved files against the NSRL and, if the hash is not in the NSRL, the file is probably interesting. For the second example, it is enough to know if a hash is in the NSRL at all. As I have this use case more than once a year, I created a workflow for this task using Redis. Redis is a simple NoSQL, key-value in-memory database. To work efficiently with the hash set, I import all SHA-1 hashes as key with the value TRUE into a Redis database. After that, I create hashes of all carved files and ask Redis if it has the hash as a key. If not, I copy the file to handle it further. You can find a script downloading the RDS Modern Minimal and importing it into Redis on GitHub.

0x3 Conclusion

The NSRL is an impressive corpus of software. It’s freely available Reference Data Sets are an invaluable resource for all the IT security specialists, data hoarders, digital archivists, and metadata nerds out there. Have fun with it and do something good!