File System Forensic Analysis

nazarijo writes "The field of investigative forensics has seen a huge surge in interest lately, with many looking to study it because of shows like CSI or the increasing coverage of computer-related crimes. Some people see a career opportunity there, and are moving toward computer forensics, marrying both law enforcement and investigations with their interest in things digital. Central to this field is the study of data storage and recovery, which requires a deep knowledge of how filesystems work. Brian Carrier's new book File System Forensic Analysis covers this topic with clarity and an uncommon skill." Read on for the rest of Nazario's review. File System Forensic Analysis author Brian Carrier pages 600 publisher Addison Wesley Professional rating 9 reviewer Jose Nazario ISBN 0321268172 summary The standard for digital filesystem forensics

It's easy to think that computer filesystems are relatively simple things. After all, if 'dir' or 'ls' don't show what you're looking for, maybe an undelete program will work. Or will it? To be a decent, trustworthy expert in forensics (a requirement if you plan to participate in any criminal investigations), you'll have to learn how filesystems really operate, how tools like undelete and lazarus work, and how they can be defeated.

Carrier's book isn't a legal book at all, and it doesn't pretend to offer much insight into the law surrounding forensics. Instead it focuses on technical matters, and is sure to be the gold standard in its field. This is important, because it comes at you expecting you to have some knowledge, even if only informal, of what a filesystem contains. With a basic understanding of data structures, you'll get a wealth of information out of this book, and it will be a good reference long after you've first studied it.

File System Forensic Analysis is divided into three sections. These are arranged in the order that you'll want to study them to maximize the benefit you can hope to achieve, namely an understanding of how to examine filesystems for hidden or previously stored data. The first three chapters cover a fundamental series of topics: Digital Investigation Foundations, Computer Foundations, and an introduction to Hard Disk Data Acquisition. While they start at a basic level (e.g. what hexadecimal is), they quickly progress to more developed topics, such as the types of interfaces (SATA, SCSI, IDE), the relationship of the disk to the computer system as a whole, and how data is stored in a file and filesystem at a basic level. A lot of examples given use Linux, due to the raw, accessible nature of UNIX and UNIX-like systems, and the availability of tools like 'dd' to gather data.

Part 2 covers "Volume Analysis," or the organization of files into a storage system. This introduces the basics of things like partition tables (including how to read one). The next few chapters cover PC-based partitions (DOS and Apple), server-based partitions (BSD, Solaris and GPT partitions), and then multiple disk volumes like RAID and logical volumes. With this introduction, the final chapter of the section covers how to use these filesystem descriptions in practice to look for data during analysis. Filesystem layouts, organization, and things like journals and consistency checks are covered with a clarity and exactness that's refreshing for such a detailed topic.

Having covered the basics of filesystems, Part 3 covers the bulk of the book and material. Several chapters follow that specifically show you how to analyze particular filesystems by using their data structures to direct your reads. A range of filesystems are covered, including FAT, NTFS, EXT2 and EXT3, and the BSD types UFS1 and UFS2. Each filesystem has two chapters, one devoted to concepts and analysis, another entirely about data structures. Dividing each filesystem type like this lets Carrier focus first on the theory of each filesystem and its design, and then the practical use of its design to actually understand how to pull data off of it.

The real strength of File System Forensic Analysis lies in Carrier's direct and clear descriptions of the concepts, the completeness of his coverage, and the detail he provides. For example, a number of clear, well-ordered and simple diagrams are peppered throughout the book, explaining everything from allocation algorithms to NTFS alternative data streams. This use of simple diagrams makes the topics more easily understood, so the book's full value can be appreciated. This is the kind of thing that sets a book apart from its peers and makes it a valuable resource for a long time.

Finally, Carrier brings it all together and shows us how many aspects of filesystems can be examined using his "sleuth kit" tools, freely available and easy to use. Without appearing to hawk this tool at the expense of other valuable resources, you get to see how simple and direct filesystem manipulations can be done using a direct approach. This kind of presentation is what makes File System Forensic Analysis a great foundation.

Overall I'm pleased with File System Forensic Analysis, I think that Carrier has achieved what few technical authors do, namely a clear explanation of highly technical topics which retains a level of detail that makes it valuable for the long term. For anyone looking seriously at electronic forensics, this is a must have. I suspect people who are working on filesystem implementations will also want to study it for its practical information about NTFS. Overall, a great technical resource.

You can purchase File System Forensic Analysis from bn.com. Slashdot welcomes readers' book reviews -- to see your own review here, read the book review guidelines, then visit the submission page.

Comments

STEP ONE

[jos3000]

Don't forget to mount the drive as read only!

CSI

[Seumas]

Why in the hell would you choose a dull career like forensic investigation based on a TV show? That would be like becoming a cop because you want to be like Dirty Harry. How many of these gits go into college for this kind of career, because they think it's going to be exciting and they're going to discover the case-cracking evidence in a few hours, grab their gun and go make an arrest?

I might get this

[L. VeGas]

This sounds really interesting. I've been fascinated for a while with how the file / folder metaphor has become so entrenched that people have a difficult time imagining any other way of thinking about it.

As the OS has become more sophisticated, most computer users now never see things like a disk defrag. They really think that there is a file, all in one spot in their computer, that sits literally next to other files in the same folder. The idea that you can recover a file that has been "deleted" seems like deep wizardry, with no thought to the more impressive wizardry that makes "files" out of pieces of metal with a magnet.

Other views on the book

[sidney]

For alternate opinions on the book see this review by Rob Slade in RISKS Digest, and this short rebuttal of Slade's review by Simson Garfinkle.

STEP ZERO:

[abb3w]

Make sure by ordering the right adapter for doing forensic's work that Your Young Apprentice (or PFY) can't screw this up. A read-only adapter means the drive can't be mounted rewritably. No, it's not cheap. But what's $500 to the assurance that your evidence chain is prevented from fuckup at the hardware level?

And no, I don't work for these people. I just think they make some nifty geek toys.

No, that's not why I have SCSI drives on my home server. Honest; it's for the RAID performance....

Re:STEP ZERO:

[Shanep]

Well, instead of using an OS that does what it damn well wants (like mount all drives read/write by default),

I agree, gathering evidence with Windows sucks.

why don't you use Linux and simply create a drive image straight from the raw device without mounting at all?

Because in court, things can get nasty like this...

Barrister: Did you use a (looks at freshly written note) "write blocker", Mr. Smith?
Forensics guy: No, I did not need to. I refrained from mounting the disk and copied it at a raw block-for-block level (confusing to judge).
Barrister: Yes or No Mr. Smith, did you use a "write blocker".
Forensics guy: No.
Barrister: And a "write blocker" is a forensics industry standard method for preventing contamination of captured evidence? (Judge respects witnesses who respect the court enough to make sure their captured evidence is absolutely accurate and original evidence could not have been altered).
Forensics guy: Yes, but...
Barrister: Mr. Smith, you failed to take a basic precaution to make absolutely certain that the captured evidence was not altered in any way, by using a basic device that is normally a part of the toolkit of a computer forensic professional. Do you posess a "write blocker" Mr. Smith?
Forensics guy: Yes (No).
Barrister: Then WHY did you not use it?! (You ARE a computer forensics professional are you not Mr. Smith?)
Forensics guy: gasp gasp (blush) choke...

The point is, if you are gathering evidence of this sort, then write blockers are tools you should have and always use. All the opposition needs to do is raise doubt. And then you and your client are screwed.

When you take the stand or put on an affidavit, the opposing legal team will attack:

1/ Your findings and the methods you used to get to them.
2/ Your evidence.
3/ You credibility.

and at a worst case...

4/ Accuse you of tampering with ORIGNAL EVIDENCE which has been tendered to the court!

Not having a write-blocker says, "I am not a computer forensics professional".

Having a write-blocker and not using it says, "I am sloppy and failed to use a simple tool at my disposal to assist the court as best I could".

Whether your evidence is exactly the same as the other forensics experts is beside the point. They have attacked your credibility and that can go against your findings (even if they are completely correct). You have nothing to gain from not using a write-blocker (which you should already have) and everything to loose. I would love to just capture evidence with FreeBSD and just copy from the raw device. But at the end of the day, the cost of a $500 write-blocker, which you get to use over and over, should be peanuts compared with what you make each day you work on cases which requires its use.

New TV show..

[Bnderan]

I will look forward to watching SCSI-Miami.

What about encryption?

[tacokill]

I know that encryption is a topic unto itself but it is becoming more and more common for people to create PGP Disks or DriveCrypt disks.

How do those things fit into this topic? I mean, the filesystem stuff is great and interesting but it doesn't seem to do any good if all you can recover is a PGP Disk file*.

Can someone much smarter than me tell me how data forensics deals with that????



* PGP Disk: a pgp encrypted file that can be mounted as a drive letter. It is, literally, a file just sitting there on your harddrive. You mount the file (after providing the secret passphrase) and voila! - you now have an encrypted drive to copy files in and out of.

Re:I do this sometimes...

[Dogtanian]

I was looking for personal ads, dating sites, child or extreme porn

What the heck is 'extreme porn'?!

People f*****g on snowboards at 120MPH? Some naked chick with massive fake breasts doing skateboard stunts on a halfpipe while guys standing at the top on each side try to bukakke her while she's paused in mid-air?

"It's not XXX rated.... it's XXXTREME rated!"

Having done forensic work...

[bradleyland]

Honestly, this job is probably the coolest I've done. We get the run of any joint we enter. We get to crack people's passwords, read their stuff, and pry into the details that they're trying to hide.

Outside of the unreal timeframe, it is a bit like television. I've been on location at 1 AM acquiring hard drives so that the debtor principles didn't know what we were doing. Walking through the data center with my mag light at that hour of the morning comes pretty close to that feeling you get when you watch CSI on TV. Most of the time, we tell the people on location we're making "backups" of the data so that we can preserve the data in the event of a crash. There's definitely a social element to forensic work (at least in bankruptcy cases).

A typical acquisition may go something like this:

You set up, pull your forms, start noting observations, pull the drives, hook them up to the little black box connected to your laptop's firewire port (a write-blocker), and start having a look at the data. If you've got what you're looking for, you acquire the drive and put everything back together. Boot it all up and be on your way.

You may be doing this in the CEO's office, or in the data center looking for a mail server. The top officers are usually the most important, since they have the most important correspondence and data.

It's a fun job. It's every bit as exciting as what you see on television (for once).