4 ms·
As long as software exists, by definition, zero-days will exist. A zero-day is simply a bug in its most nascent state; one person has found it, and nobody else
by chatmasta 11y ago
As long as software exists, by definition, zero-days will exist. A zero-day is simply a bug in its most nascent state; one person has found it, and nobody else knows about it. Whether the finder is a "security researcher," a "blackhat," or a "nation-state" has no impact on whether the bug exists or not. In fact, the bug exists even if nobody finds it! The distinction of who found it, and what they do with it, is purely political. Anyone can still exploit the bug.
Sure, maybe the "friendlier" bug finders will responsibly disclose any bugs they find. But there will never be a way to guarantee that all bugs found will be responsibly disclosed. Even if we convince the FBI/NSA to "responsibly disclose" every bug they find (will never happen), what about every other country? The hundreds of security firms? The thousands of independent hackers and "researchers?"
Zero-days will ALWAYS exist. Software will ALWAYS be exploitable. Worrying about how people react when they find those exploits is the similar to arguing about gun control. Sure, maybe we can convince some actors to responsibly disclose, but the bad actors will always keep the exploits for themselves and use them "irresponsibly." And there will always be bad actors.
So instead of fretting about what happens when someone finds a bug, why don't we prepare for the eventuality that all bugs will be found and exploited, often times without anyone's knowledge? Why don't we build security systems to be tolerant of exploits, instead of resistant to them? There is no security panacea, just as there is no reliability panacea.
We build distributed systems with the assumption that nodes will fail, and we call that "fault tolerance." We don't say a system is broken because a node fails. We say it's broken if it cannot handle a node failing.
Why can't we do the same for our security systems? Exploits are as inevitable as any type of system failure. We need to design for exploit tolerance with the same enthusiasm we design for fault tolerance.
- ccvannorman 11y agoWow - well said. I have often said that modern computer security is akin to trying to float on the ocean in a sieve while constantly plugging each hole in the sieve all the time. I think the world is ready for a new kind of computer. Want to start a company? ;-]
- nickpsecurity 11y agoLook up CHERI processor and CHERIBSD at Cambridge. Look up Gaisler Leon3/4 SOC. License them, port CHERI tech to Leon, add capability-aware IOMMU, add TRNG, add ROM for main bootloader, fab the joker, and you got critical part of a secure computer. Clone Tandem Nonstop lockstep tech if you want fault-tolerance. And there you go. A secure PC for only a few million dollars with BSD support. Optionally port EROS microkernel w/ jVPS filesystem for stronger solution.
- sqeaky 11y agoExploits being inevitable does not exonerate our elected and appointed officials from their responsibilities and obligation of running the government the people who put them in power want. Do we want a government that collects threats that are more likely to be used against its own people than anyone else? The USA produces and use more software than any other country, not sharing exploit information fundamentally hurts the USA and the Economy of the USA more than other countries.
- tptacek 11y agoIt's a pretty big leap you've taken here, between your disagreement with the USG's policy regarding vulnerabilities and their "obligation to run the government people want". Clearly, they're not running the government that message boards want. But that's not the same thing!
- sqeaky 11y agoI remember some old document or preseident saying saying something about a "government for the people, by the people". I suspect Google, Microsoft and most businesses involving the Internet and Money at the same time care. They probably want to know about flaws so they don't wind up losing their pants like the Bangladesh bank did recently. There is huge money and real lives on the line here, this is not about appeasing message board hackers. Imagine if some big hack was learned to be known about in a 0-day the FBI withheld? I am fairly certain it was Target's mismanagement, but imagine if the a document was leaked and it showed the government knew about a 0-day that allowed the Target hack. That is small compared to what might actually happened when talking about browser or OS attack surface.
- tptacek 11y agoJust because you personally disagree with something the government is doing doesn't mean they aren't doing it "for the people". You have to compose a real argument.