4 ms·
>It's just about impossible for an outsider to reverse engineer a hard drive, read the existing firmware, and create malicious versions. It's hard, but it isn'
by wongarsu 12y ago
>It's just about impossible for an outsider to reverse engineer a hard drive, read the existing firmware, and create malicious versions.
It's hard, but it isn't like this hasn't been done before: http://spritesmods.com/?art=hddhack&page=3 http://spritesmods.com/?art=hddhack&page=3
And to be honest, I think what they call "about impossible" is just a question of investing enough time and having a somewhat decent understanding of microprocessors and reverse-engineering. I'm not trying to downplay Equation Group's achievement, but it feels like arstechnica is starting to exaggerate here.
- damurdock 12y agoThe whole "embedded programming is scary and impossible" sentiment seems to be pretty common these days. I mean, it's impressive, but well within the reach of an interested nation-state or even a skilled hobbyist like Sprite_tm. It seems like the prevalence of modified router firmwares (and botnets based on them) should make it clear that firmware modification is A Thing That Happens(TM).
- tptacek 12y agoAnyone who thinks this kind of work requires nation-state backing should look at the (hobbyist!) projects to jailbreak the Xbox, X360 and PS3. Low-level kernel- and hypervisor- programming in which obstacles are casually overcome by dropping zero-day memory corruption bugs in core libraries that would be worth tens of thousands of dollars today just from bounty programs, deploying crypto bugs that make "custom RC5" and "thousands of iterations of SHA1" look like the shoplifted- from- Schneier technique that it appears to actually be (TEA hash collision, ECDSA nonce repeats, &c). And these were, more or less, student side projects.
- venomsnake 12y agoPiracy and cheating has always been gateway to serious programming.
- trvz 12y agoWait, what? "Piracy and cheating" isn't always a gateway to serious programming. Exploring the software on your purchased devices can be a gateway to serious programming. What you call that process is subjective, but arguably not piracy and cheating.
- throwaway49873 12y agopiracy = cracking. cheating = trainers i.e. game cheats, which is very similar.
- javajosh 12y agoThat doesn't mean that it's easy, it just means that hacking a PS3 is highly motivating. Making some lame centerfuge explode half-way around the world is high on patriotism (I guess) but short on lols.
- tptacek 12y agoIt's not easy (the sophistication of some of the console jailbreaking work is extremely frustrating to some of us in the pro vuln research community, in that they treat as footnotes things that would score a whole Black Hat talk).
- wongarsu 12y agoHe didn't say that it's easy, just that it isn't so hard that you need nation-state backing to do it. Judging purely from their capabilities, Equation Group could be a small group of college graduates who make a living with hacking and don't settle for the low-hanging fruit. Of course the other evidence strongly suggests that this group is NSA-employed, but the real reason why their malware is so much more sophisticated than anything else out there isn't that nation states have staggering amounts of resources, it's simply that normal malware doesn't need to be sophisticated.
- cocoablazing 12y agoYou are saying that you think it is reasonable that a small group of "college graduates" have created and executed a global espionage campaign against airgapped military targets using software that makes previous nec plus ultra NSA cyberweapons look obsolete.
- hengheng 12y agoThere's still a difference between finding the weakest link in a big chain and then attacking into it, and creating large amounts of features out of thin air. The latter is not any more difficult, but it requires one or two magnitudes more effort, which is almost impossible without a larger team. I'm thinking 2 people in a garage vs 30-60 people in an office building.
- lawnchair_larry 12y agoNot at all similar. This absolutely requires nation state budgets.
- tptacek 12y ago... because...
- lawnchair_larry 12y agoIt's worth noting that those cases used open source firmware.
- zaroth 12y agoI remember reading this on HN a while back and came here to post it as well. Techniques like this show funding and/or determination but not really brilliance. If someone asked you to permanently infect hardware, the hard drive controller might be the first place you look, and certainly not the last. The interdiction seems to be the strongest indicator of funding level. The so-called Equation Group doesn't sound like much more than a group of average engineers doing their job. Breaking into systems is mostly childs play. Securing them is the more interesting challenge.
- comboy 12y agoIn light of this thread, there's an interesting line from the article you've linked: > The Cortex-M3 handles... nothing? I could stop it and still have all hard disk functions. I mean, hacking is fun, but being a government I guess you can also convince some manufacturers to make things easier.
- MichaelGG 12y agoIt's idiocy to think "outsiders" are needed in the first place. The CIA ACD FBI had multi decade breaches that cost the Soviets/Russia less than a couple million dollars. If FBI agents can be flipped for so little, it's silly to think the NSA cannot get detailed engineering plans and source code for every electronics product they desire. Not to mention they can always recruit people to work totally undercover in industry. Hire top performers, give them coaching or even sabotage rival colleagues. It certainly cannot cost more than a few million per agent to deeply infiltrate target companies. Hell, in their place, I'd be trying to get multiple people into MS, Intel, etc. to be in position to use root keys if needed.
- lawnchair_larry 12y agoThat statement isn't from ars, it's from experts, and it's correct. Reverse engineer 12+ firmwares, make extensive modifications to their functionality, recompile them, properly QA them, and use them as just one element of a complex malware chain is not a student project. The link you provided shows how much investment it takes to identify a small bit of firmware functionality and not even do anything useful with it.