Schneier on Security

benEzra • January 2, 2014 10:44 PM

@3ik4fbh,

“I’m guessing you don’t know every tier of law enforcement in the US have almost unregulated access to firearms manufacturing databases with unique rifling geometry per model.. It takes a matter of minutes to track firearm sales and ownership with this data”

Barrel markings on a bullet are like tire tracks, not fingerprints. Given an intact bullet (not impossible, though not certain), one can usually tell the caliber, usually infer the chambering (e.g. .22LR vs. .223 vs. .22 Hornet vs. .22-250 if it’s a .22 caliber bullet), and over short time scales you can match the individual gun that fired it from a small sample of suspect guns. What you can’t do is match bullet X to an individual gun that has been shot a lot since the ballistic sample was taken, or pick a gun out of a sample size of thousands or millions. Just as tire tracks can tell you that “a sports car, possibly a late-model Corvette or CTS, shod with fairly new Michelin Pilot Sport tires, was here”, and if you have a couple dozen suspects, you can tell that those tracks probably came from the Goodyears on Suspect 7’s 2012 CTS-V, but you cannot examine tire tracks and pick a car out of all Vettes and CTS’s in a region.

I don’t know what these “military style” rifles were (and the term is so broad as to encompass most firearms, from a Brown Bess to a Remington 700), but assuming the reporter means “chambered in a military caliber”, they could be anything from smallish .223/5.56x45mm or 7.62x39mm, to more powerful rounds like like .308 Winchester/7.62x51mm NATO, .30-06 Springfield, .300 Winchester Magnum, or .50 BMG. Thing is, the first four in that list are four of the most common civilian rifle calibers in the USA, so caliber/make/model doesn’t really tell you much if it was one of those.

Nor do I know what caliber rifle was used, but .223/5.56mm (as kicked around upthread) would seem an odd choice if the culprits wanted to actually break things, since it is the least powerful of common centerfire calibers.

@Dry Run,

“150 shots were fired in about a minute, which indicates some sort of automatic weapon. That is why they used the term “military” attack. Also, the shots were fired at night under poor lighting conditions. It is likely that they used infrared goggles to see the hot transformer radiators. This could be why they didn’t shoot the one that was off.”

As has been mentioned upthread, 150 shots in a minute or two, from two people, is ~1 round per second or so, entirely do-able with a small-caliber non-automatic rifle. If it was actually two minutes, that’s more like a shot every two seconds, which is rather slow and deliberate.

And unless it was pitch dark, transformers are not exactly hard to see even without flashlights. We probably can’t infer sophisticated equipment without further info.

@bitmonki,

“I’ve never used an AR-15, so don’t know if it has burst mode, but I think it would be a fairly trivial customization for a competent gunsmith.”

All post-1986 civilian AR-15’s are designed to be as difficult to convert to auto or burst as any other civilian self-loading rifle; they use civilian-only lower receivers and are designed so that burst or auto fire-control parts won’t fit.

benEzra • January 3, 2014 7:31 AM

@Anura,

“According to this it was 7.62×39 rounds:

http://www.texasre.org/Lists/Calendar/Attachments/605/Item%204d%20-%20NERC%20CIPC%20Report%20to%20TRE%20MRC%20-%202013Jun14.pdf”

Thanks for the link; that sheds some light, particularly the detailed knowledge of the installation that it demonstrates. Interesting choice of a relatively low-powered round, fired low on the radiators to maximize coolant loss from the holes, and the fact that only energized transformers were targeted (because coolant loss doesn’t hurt a turned-off transformer); someone unfamiliar with power systems would have been more likely to fire center-mass at all of them, and the knowledge of the communications lines that were cut may also suggest inside knowledge.

Based on that article, I suspect the “vandalism by disgruntled current/former employee” angle is more likely than the “terrah” angle, whether domestic or otherwise.

benEzra • January 3, 2014 4:36 PM

@3ik4fbh,

“Manufacturers are required by federal US law, when manufacturing inside the country, to turn in rifling data per-model. Nanometer scale is used, sometimes across models, for the rifling machining and data. Feel free to ask someone who has internal knowledge of gun manufacturing to verify; it’s actually a federal felony to not do so.”

Citation, please? According to the Gun Control Act of 1968 as amended, the controlled and tracked portion of the firearm is the receiver (or the frame, in the case of handguns), not the barrel, and often the barrel is not even manufactured by the gunmaker. (Case in point, my competition rifle is manufactured by Rock River Arms but has a Wilson barrel.) And no rifle barrel on this planet is manufactured to nanometer tolerances.

You are correct that firearm make/model can often be inferred from the rifling design, land shape, groove depth, and pitch/twist rate (just as the identity of a car can be narrowed down by tire size, track width, and wheelbase if available), dimensions which I’m sure LE agencies have access to. In the case of a popular caliber like a 7.62x39mm, that may get you down to a few tens of thousands to hundreds of thousands of units. Or not, depending on model and how individualized the specs are; I’d expect that 7.62x39mm barrels as used in this incident tend to be less individualized than, say, .223/5.56mm or .308 chambered barrels, and Eastern European barrels tend to have rougher tolerances than most Western European or American barrels.

“This data is easily correlated to track gun sales; of course not down to an indevidual gun as I thought I made clear in the first place. Which are also regulated with data-entry legislature, and have been for a quarter of a century except for some rifles and shotgun classes.”

You said that examination of the bullets would, given competence, identify the culprits, not merely provide broad caliber/model information, and I’m not seeing how you could even come close to that level of specificity.

“It’s not near the fiction as you and others here suggest. Neither are symmetry algorithms used to recover partial image data and human symmetry; there are actually software solutions used by government and civilians that use them with significant success rates.”

Image processing software can do remarkable things, but it cannot identify an individual rifle S/N from a recovered bullet.